- ieeefp-tests-cvs@common-lisp.net - mailman3.common-lisp.net

[ieeefp-tests-cvs] Is Israel a democracy? -- The problem with intellectually insecure whites -- Should Christians Support Israeli Terrorism in Gaza?
by Lawrence Auster 23 Jan '09

23 Jan '09

The Jewish State of Israel has no constitution, nor does it name its borders. Israel's hidden constitution is Judaism. Israel's undeclared borders range from the Nile to the Euphrates rivers. Israel's desired jurisdiction extends over the entire Earth. It could not be more clear that the Jewish State follows a foreign policy which obeys Jewish Law as iterated in the Hebrew Bible, the Talmud, Maimonedes, the Cabalah, and the many commentaries and refinements of same. The Jews are genociding the native inhabitants of Palestine, just as their religion advises, and because their religion teaches them to do so. They treat non-Jews as if non-humans, just as their religion requires them to do. They make perpetual war on every nation on Earth, just as their genocidal Jewish God has instructed. The Jews of Israel are simply being Jews. Jews are an existential threat to the human race. Israel contains one third of the Jews of the World. It is not some aberration of the Jewish spirit, but the condensation and concentration of the perverse Jewish mentality, which malady also pervades the remaining two thirds of Jewry, who almost unanimously support the Jewish State, and who certainly do unanimously support the Jewish People and its consistent and constant crimes against the human race. Israel is Jewry and the danger of Israel is the danger of the Jewish People to all others, as the Jews have demonstrated each and every day of their existence. The Jews, the entire Jewish People of 15 million, will not relent until they have wiped out all non-Jews in "Greater Israel". They will not stop destroying all other cultures, nations, religions, ethnicities, races, competition, etc. until they are either stopped, or succeed in their ancient quest to destroy the human race. What Israel is doing is not some reaction to outside forces, nor was the formation of Israel a response to the Holocaust. Israel is simply following the plan laid out in the Jews' religious texts. The Jews have openly planned to take Palestine and genocide the native population of Palestine for some 2,500 years before the Holocaust. The Jews have openly complained that "anti-Semitism" is a threat that gives them the right to genocide the Palestinians, not merely since the advent of Nazism, but for some 2,500 years. The Jewish religion is the Constitution of the Jewish State of Israel, and, to a greater or lesser extent, the constitution of the nature of every Jew alive. The borders of Israel are the range the Jew roams over the entire World. The perverse Jewish mentality is inbred by a Jew's exposure to his parents and to his community. Judaism passes in the spit and slobber of Jewish mother telling her Jewish child that he is a "Jew", as much as Judaism passes in the poison and pain of a Talmudic tractate. The secular Jews did not suddenly come to life after the Enlightenment and the Jewish Reformation a body of vampires that appeared ex nihilo, in vacuo, mostly atheistical and undetached from formally practiced Judaism. Judaism is the Jew. It is a mindset that transcends and supercedes religion. It is a belief set, a way of life, a perception of one's self and one's relation to the World that makes a Jew, a Jew, and a danger to all of humanity. In fact, the religious shell of Judaism is like the stretched and infected skin of a lycanthropic pustule. When you lance it to cure the infection, the virus only becomes more contagious and spills directly on the non-Jew. The secular Jew is a deliberate product of the hyper-religious Jew, a monster created out of the hewed corpses of the fanatically religious Jew, a Golem which is conjured up to enter the World of the non-Jew and poison its blood, and boil its brain with a rabid lunacy that bites and spreads, until the infected community feeds on itself and fills the fields with rotting bloating bodies, where once human beings tilled the soil and tended to their families. The religious Jew created the secular Jew as an army of Esthers who seduce with open thighs, broad smiles, and a Siren call that lures in the non-Jew to cast his skull upon the jagged rocks and color the seas with his blood, sickened and blinded by the venereal disease of Judaism in secular form. Israel is not a secular democracy. It is a religious mockery. It is a rabid bat flying to the ends of the Earth, to end the Earth. No one will be free nor safe until the disease is quarantined and dies out. Source: http://www.ziopedia.org/articles/israel/how_can_israel_claim_to_be_a_%27dem… -------------------- The problem with intellectually insecure whites By Kevin MacDonald January 19, 2009 America will soon have a white minority. This is a much desired state of affairs for the hostile elites who hold political power and shape public opinion. But it certainly creates some management issues � at least in the long run. After all, it�s difficult to come up with an historical example of a nation with a solid ethnic majority (90% white in 1950) that has voluntarily decided to cede political and cultural power. Such transformations are typically accomplished by military invasions, great battles, and untold suffering. And it�s not as if everyone is doing it. Only Western nations view their own demographic and cultural eclipse as a moral imperative. Indeed, as I have noted previously, it is striking that racial nationalism has triumphed in Israel at the same time that the Jewish intellectual and political movements and the organized Jewish community have been the most active and effective force for a non-white America. Indeed, a poll in 2008 found that Avigdor Lieberman was the second most popular politician in Israel. Lieberman has advocated expulsion of Arabs from Israel and has declared himself a follower of Vladimir Jabotinsky, the leading pioneer of racial Zionism. The most popular politician in the poll was Benjamin Netanyahu � another admirer of Jabotinsky. Prime Minister Ehud Olmert and Foreign Minister Tzipi Livni are also Jabotinskyists. The racial Zionists are now carrying out yet another orgy of mass murder after a starvation-inducing blockade and the usual triggering assault designed to provoke Palestinian retaliation � which then becomes the cover for claims that Israel is merely defending itself against terrorism. This monstrosity was approved by overwhelming majorities of both Houses of Congress. The craven Bush administration did its part by abstaining from a UN resolution designed by the US Secretary of State as a result of a personal appeal by the Israeli Prime Minister. This is yet another accomplishment of the Israel Lobby, but one they would rather not have discussed in public. People might get the impression that the Lobby really does dictate US foreign policy in the Mideast. Obviously, such thoughts are only entertained by anti-Semites. But I digress. In managing the eclipse of white America, one strategy of the mainstream media is to simply ignore the issue. Christopher Donovan (�For the media, the less whites think about their coming minority status, the better�) has noted that the media, and in particular, the New York Times, are quite uninterested in doing stories that discuss what white people think about this state of affairs. It�s not surprising that the New York Times � the Jewish-owned flagship of anti-white, pro-multicultural media � ignores the issue. The issue is also missing from so-called conservative media even though one would think that conservatives would find the eclipse of white America to be an important issue. Certainly, their audiences would find it interesting. Now we have an article �The End of White America� written by Hua Hsu, an Assistant Professor of English at Vassar College. The article is a rather depressing display of what passes for intellectual discourse on the most important question confronting white people in America. Hsu begins by quoting a passage in F. Scott Fitzgerald�s The Great Gatsby in which a character, Tom Buchanan, states: �Have you read The Rise of the Colored Empires by this man Goddard?� � Well, it�s a fine book, and everybody ought to read it. The idea is if we don�t look out the white race will be�will be utterly submerged. It�s all scientific stuff; it�s been proved.� Buchanan�s comment is a thinly veiled reference to Lothrop Stoddard�s The Rising Tide of Color which Hsu describes as �rationalized hatred� presented in a scholarly, gentlemanly, and scientific tone. (This wording that will certainly help him when he comes up for tenure.) As Hsu notes, Stoddard had a doctorate from Harvard and was a member of many academic associations. His book was published by a major publisher. It was therefore �precisely the kind of book that a 1920s man of Buchanan�s profile � wealthy, Ivy League�educated, at once pretentious and intellectually insecure � might have been expected to bring up in casual conversation.� Let�s ponder that a bit. The simple reality is that in the year 2009 an Ivy League-educated person, "at once pretentious and intellectually insecure," would just as glibly assert the same sort of nonsense as Hsu. To wit: The coming white minority does not mean that the racial hierarchy of American culture will suddenly become inverted, as in 1995�s White Man�s Burden, an awful thought experiment of a film, starring John Travolta, that envisions an upside-down world in which whites are subjugated to their high-class black oppressors. There will be dislocations and resentments along the way, but the demographic shifts of the next 40 years are likely to reduce the power of racial hierarchies over everyone�s lives, producing a culture that�s more likely than any before to treat its inhabitants as individuals, rather than members of a caste or identity group. The fact is that no one can say for certain what multicultural America without a white majority will be like. There is no scientific or historical basis for claims like �the demographic shifts of the next 40 years are likely to reduce the power of racial hierarchies over everyone�s lives, producing a culture that�s more likely than any before to treat its inhabitants as individuals, rather than members of a caste or identity group.� Indeed, there is no evidence at all that we are proceeding to a color blind future. The election results continue to show that white people are coalescing in the Republican Party, while the Democrats are increasingly the party of a non-white soon-to-be majority. Is it so hard to believe that when this coalition achieves a majority that it will further compromise the interests of whites far beyond contemporary concerns such as immigration policy and affirmative action? Hsu anticipates a colorblind world, but affirmative action means that blacks and other minorities are certainly not treated as individuals. And it means that whites � especially white males � are losing out on opportunities they would have had without these policies and without the massive non-white immigration of the last few decades. Given the intractability of changing intelligence and other traits required for success in the contemporary economy, it is unlikely that 40 more years of affirmative action will attain the outcomes desired by the minority lobbies. Indeed, in Obama's America, blacks are rioting in Oakland over perceived racial injustices, and from 2002 �2007, black juvenile homicide victims increased 31%, while black juvenile homicide perpetrators increased 43%. Hence, the reasonable outlook is for a continuing need for affirmative action and for racial activism in these groups, even after whites become a minority. Whites will also lose out because of large-scale importation of relatively talented immigrants from East Asia. Indeed, as I noted over a decade ago, "The United States is well on the road to being dominated by an Asian technocratic elite and a Jewish business, professional, and media elite." Hsu shows that there already is considerable anxiety among whites about the future. An advertizing executive says, �I think white people feel like they�re under siege right now � like it�s not okay to be white right now, especially if you�re a white male. ... People are stressed out about it. �We used to be in control! We�re losing control�� Another says, "There�s a lot of fear and a lot of resentment." It's hard to see why these feelings won't increase in the future. A huge problem for white people is lack of intellectual and cultural confidence. Hsu quotes Christian (Stuff White People Like) Lander saying, "I get it: as a straight white male, I�m the worst thing on Earth." A professor comments that for his students "to be white is to be culturally broke. The classic thing white students say when you ask them to talk about who they are is, �I don�t have a culture.� They might be privileged, they might be loaded socioeconomically, but they feel bankrupt when it comes to culture � They feel disadvantaged, and they feel marginalized." This lack of cultural confidence is no accident. For nearly 100 years whites have been subjected to a culture of critique emanating from the most prestigious academic and media institutions. And, as Hsu points out, the most vibrant and influential aspect of American popular culture is hip-hop�a product of the African American urban culture. The only significant group of white people with any cultural confidence centers itself around country music, NASCAR, and the small town values of traditional white America. For this group of whites � and only this group � there is "a racial pride that dares not speak its name, and that defines itself through cultural cues instead�a suspicion of intellectual elites and city dwellers, a preference for folksiness and plainness of speech (whether real or feigned), and the association of a working-class white minority with 'the real America.'� This is what I term implicit whiteness � implicit because explicit assertions of white identity have been banned by the anti-white elites that dominate our politics and culture. It is a culture that, as Hsu notes, "cannot speak its name." But that implies that the submerged white identity of the white working class and the lack of cultural confidence exhibited by the rest of white America are imposed from outside. Although there may well be characteristics of whites that facilitate this process, this suppression of white identity and interests is certainly not the natural outcome of modernization or any other force internal to whites as a people. In my opinion, it is the result of the successful erection of a culture of critique in the West dominated by Jewish intellectual and political movements. The result is that educated, intellectually insecure white people these days are far more likely to believe in the utopian future described by Hsu than in hard and cautious thinking about what the future might have in store for them. It's worth dwelling a bit on the intellectual insecurity of the whites who mindlessly utter the mantras of multiculturalism that they have soaked up from the school system and from the media. Most people do not have much confidence in their intellectual ability and look to elite opinion to shape their beliefs. As I noted elsewhere, A critical component of the success of the culture of critique is that it achieved control of the most prestigious and influential institutions of the West, and it became a consensus among the elites, Jewish and non-Jewish alike. Once this happened, it is not surprising that this culture became widely accepted among people of very different levels of education and among people of different social classes. Most people are quite insecure about their intellectual ability. But they know that the professors at Harvard, and the editorial page of the New York Times and the Washington Post, and even conservative commentators like Rush Limbaugh and Sean Hannity are all on page when it comes to racial and ethnic issues. This is a formidable array, to the point that you almost have to be a crank to dissent from this consensus. I think one of the greatest triumphs of the left has been to get people to believe that people who assert white identity and interests or who make unflattering portrayals of organized Jewish movements are morally degenerate, stupid, and perhaps psychiatrically disturbed. Obviously, all of these adjectives designate low status. The reality is that the multicultural emperor has no clothes and, because of its support for racial Zionism and the racialism of ethnic minorities in America, it is massively hypocritical to boot. The New York Times, the academic left, and the faux conservatives that dominate elite discourse on race and ethnicity are intellectually bankrupt and can only remain in power by ruthlessly suppressing or ignoring the scientific findings. This is particularly a problem for college-educated whites. Like Fitzgerald's Tom Buchanan, such people have a strong need to feel that their ideas are respectable and part of the mainstream. But the respectable mainstream gives them absolutely nothing with which to validate themselves except perhaps the idea that the world will be a better place when people like them no longer have power. Hsu quotes the pathetic Christian Lander: "�Like, I�m aware of all the horrible crimes that my demographic has done in the world. ... And there�s a bunch of white people who are desperate � desperate � to say, �You know what? My skin�s white, but I�m not one of the white people who�s destroying the world.�� As a zombie leftist during the 1960s and 1970s, I know what that feeling of desperation is like � what it's like to be a self-hating white. We must get to the point where college-educated whites proudly and confidently say they are white and that they do not want to become a minority in America. This reminds me of the recent docudrama Milk, which depicts the life of gay activist Harvey Milk. Milk is sure be nominated for an Oscar as Best Picture because it lovingly illustrates a triumph of the cultural left. But is has an important message that should resonate with the millions of whites who have been deprived of their confidence and their culture: Be explicit. Just as Harvey Milk advocated being openly gay even in the face of dire consequences, whites need to tell their family and their friends that they have an identity as a white person and believe that whites have legitimate interests as white people. They must accept the consequences when they are harassed, fired from their jobs, or put in prison for such beliefs. They must run for political office as openly pro-white. Milk shows that homosexuals were fired from their jobs and arrested for congregating in public. Now it's the Southern Poverty Law Center and the rest of the leftist intellectual and political establishment that harasses and attempts to get people fired. But it's the same situation with the roles reversed. No revolution was ever accomplished without some martyrs. The revolution that restores the legitimacy of white identity and the legitimacy of white interests will be no exception. But it is a revolution that is absolutely necessary. The white majority is foolish indeed to entrust its future to a utopian hope that racial and ethnic identifications will disappear and that they won�t continue to influence public policy in ways that compromise the interests of whites. It does not take an overactive imagination to see that coalitions of minority groups could compromise the interests of formerly dominant whites. We already see numerous examples in which coalitions of minority groups attempt to influence public policy, including immigration policy, against the interests of the whites. Placing ourselves in a position of vulnerability would be extremely risky, given the deep sense of historical grievance fostered by many ethnic activists and organized ethnic lobbies. This is especially the case with Jews. Jewish organisations have been unanimous in condemning Western societies, Western traditions, and Christianity, for past crimes against Jews. Similar sentiments are typical of a great many African Americans and Latinos, and especially among the ethnic activists from these groups. The �God damn America� sermon by President Obama's pastor comes to mind as a recent notorious example. The precedent of the early decades of the Soviet Union should give pause to anyone who believes that surrendering ethnic hegemony does not carry risks. The Bolshevik revolution had a pronounced ethnic angle: To a very great extent, Jews and other non-Russians ruled over the Russian people, with disastrous consequences for the Russians and other ethnic groups that were not able to become part of the power structure. Jews formed a hostile elite within this power structure � as they will in the future white-minority America; Jews were �Stalin�s willing executioners.� Two passages from my review of Yuri Slezkine's The Jewish Century seem particularly appropriate here. The first passage reminds me of the many American Jews who adopt a veneer of support for leftist versions of social justice and racial tolerance while nevertheless managing to support racial Zionism and the mass murder, torture, and incarceration of the Palestinian people in one of the largest prison systems the world has ever seen. Such people may be very different when they become a hostile elite in a white-minority America. Many of the commentators on Jewish Bolsheviks noted the �transformation� of Jews [after the Bolshevik Revolution]. In the words of [a] Jewish commentator, G. A. Landau, �cruelty, sadism, and violence had seemed alien to a nation so far removed from physical activity.� And another Jewish commentator, Ia. A. Bromberg, noted that: the formerly oppressed lover of liberty had turned into a tyrant of �unheard-of-despotic arbitrariness��. The convinced and unconditional opponent of the death penalty not just for political crimes but for the most heinous offenses, who could not, as it were, watch a chicken being killed, has been transformed outwardly into a leather-clad person with a revolver and, in fact, lost all human likeness. ... After the Revolution, ... there was active suppression of any remnants of the older order and their descendants. ... The mass murder of peasants and nationalists was combined with the systematic exclusion of the previously existing non-Jewish middle class. The wife of a Leningrad University professor noted, �in all the institutions, only workers and Israelites are admitted; the life of the intelligentsia is very hard� (p. 243). Even at the end of the 1930s, prior to the Russification that accompanied World War II, �the Russian Federation�was still doing penance for its imperial past while also serving as an example of an ethnicity-free society� (p. 276). While all other nationalities, including Jews, were allowed and encouraged to keep their ethnic identities, the revolution remained an anti-majoritarian movement. The difference from the Soviet Union may well be that in white-minority America it will not be workers and Israelites who are favored, but non-whites and Israelites. Whites may dream that they are entering the post-racial utopia imagined by their erstwhile intellectual superiors. But it is quite possible that they are entering into a racial dystopia of unimaginable cruelty in which whites will be systematically excluded in favor of the new elites recruited from the soon-to-be majority. It's happened before. Kevin MacDonald is a professor of psychology at California State University�Long Beach. Permanent URL with hyperlinks: http://www.theoccidentalobserver.net/articles/MacDonald-Hsu.html ----------- Should Christians Support Israeli Terrorism in Gaza? A timely discussion between Rev. Ted Pike and Dr. David Duke, one especially important for the Christians in our audience http://www.davidduke.com/mp3/dukeradio090122DukeandPikeonGaza.mp3 In this vital discussion, Rev. Pike and Dr. Duke explore the Pro-Israel attitude of some Christian evangelical organizations, and why their position not only goes directly against Christian morality and decency, but actually is directly opposite of that expressed by Christian Scriptures. Today, Many Christians are instructed that Jews and today�s Israel has a special covenant� with God. In fact, the New Testament in the clearest of language states that the Jews �continued not in my covenant, and I considered them not, saith the Lord.� Here�s the quote that Christians aren�t supposed to notice.: 8:10 Not according to the covenant that I made with their fathers, in the day when I took them by the hand out of the land of Egypt; because they continued not in my covenant, and I regarded them not, saith the Lord. (Hebrews 8:10) They also don�t seem to notice that a 2000 year old Judaic war against Christianity that has been waged since time of Jesus Christ and still goes on today with the most powerful Jewish organizations attempting to destroy European and American traditions, that has even become a war on our Christmas traditions. Dr. Duke and Ted Pike also speak about how over a hundred thousand Christian Palestinians have suffered with their families from anti-Christian Israel! Christian support of Israel has resulted in the very birthplace of Jesus Christ, go from 90 percent Palestinian Christians to 35 percent today because of Israeli terror and occupation. They ask, �How could any Christian in good conscience support the anti-Christian state of Israel, bombing the homes, killing and maiming, torturing and oppressing fellow Christian men, women and children?� This is a vital show for every Christian reader and listener of DavidDuke.com. Next time, you hear someone say, �God tells us that we must support Israel� you will have the clear Christian answer that just the opposite is true! For documentation on this be sure to read some of the well-footnoted, sample chapters of Jewish Supremacism and My Awakening. Source : http://www.davidduke.com/general/should-christians-support-israeli-terroris… ------------------------------------- You or someone using your email adress is currently subscribed to the Lawrence Auster Newletter. If you wish to unsubscribe from our mailing list, please let us know by calling to 1 212 865 1284 Thanks, Lawrence Auster, 238 W 101 St Apt. 3B New York, NY 10025 Contact: lawrence.auster(a)att.net -------------------------------------

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/Notes
by rtoy＠common-lisp.net 25 Aug '05

25 Aug '05

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests In directory common-lisp.net:/tmp/cvs-serv15937 Added Files: Notes Log Message: Initial revision. Date: Thu Aug 25 20:05:25 2005 Author: rtoy

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieee754/ieee754-cmucl.lisp
by rtoy＠common-lisp.net 25 Aug '05

25 Aug '05

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ieee754 In directory common-lisp.net:/tmp/cvs-serv15864/ieee754 Modified Files: ieee754-cmucl.lisp Log Message: Cargo cult the macrolet from ieee754-sbcl.lisp to get ieee754:= defined. Date: Thu Aug 25 20:04:17 2005 Author: rtoy Index: ieeefp-tests/ieee754/ieee754-cmucl.lisp diff -u ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.4 ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.5 --- ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.4 Thu Aug 25 18:43:41 2005 +++ ieeefp-tests/ieee754/ieee754-cmucl.lisp Thu Aug 25 20:04:17 2005 @@ -108,3 +108,23 @@ (incf f) (decf f)) (* s (scale-float (float f x) e)))))) + +(macrolet + ((def (x &body body) + `(defun ,x (x y) + (declare (type float x y)) + ,@body))) + (def = (cl:= x y)) + (def ?<> (not (= x y))) + (def > (cl:> x y)) + (def >= (cl:>= x y)) + (def < (cl:< x y)) + (def <= (cl:<= x y)) + (def ? (or (ext:float-nan-p x) (ext:float-nan-p y))) + (def <> (or (< x y) (> x y))) + (def <=> (or (< x y) (= x y) (> x y))) + (def ?> (or (? x y) (> x y))) + (def ?>= (or (? x y) (>= x y))) + (def ?< (or (? x y) (< x y))) + (def ?<= (or (? x y) (<= x y))) + (def ?= (or (? x y) (= x y))))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieeefp-tests.lisp
by rtoy＠common-lisp.net 25 Aug '05

25 Aug '05

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests In directory common-lisp.net:/tmp/cvs-serv10364 Modified Files: ieeefp-tests.lisp Log Message: Modify the :eq tests so that it's easier to tell what's happening. The expected result is the actual bit pattern. We also try to replace the actual result with one of the special values to make it easier to see what's happening too. Date: Thu Aug 25 18:49:24 2005 Author: rtoy Index: ieeefp-tests/ieeefp-tests.lisp diff -u ieeefp-tests/ieeefp-tests.lisp:1.11 ieeefp-tests/ieeefp-tests.lisp:1.12 --- ieeefp-tests/ieeefp-tests.lisp:1.11 Mon Aug 2 14:54:12 2004 +++ ieeefp-tests/ieeefp-tests.lisp Thu Aug 25 18:49:17 2005 @@ -203,7 +203,7 @@ (defun make-double-result-test-form (vector) `(let ((result-bits (double-float-bits result))) ,(ecase (test vector) - (:eq `(= result-bits ,(expected-answer vector))) + (:eq '(maybe-replace-special-value result-bits)) (:uo `(= (logand +quiet-double-float-nan-mask+ result-bits) +quiet-double-float-nan-mask+)) ((:vn :nb) @@ -236,7 +236,7 @@ (defun make-single-result-test-form (vector) `(let ((result-bits (single-float-bits result))) ,(ecase (test vector) - (:eq `(= result-bits ,(expected-answer vector))) + (:eq '(maybe-replace-special-value result-bits)) (:uo `(= (logand result-bits +quiet-single-float-nan-mask+) +quiet-single-float-nan-mask+)) ((:vn :nb) @@ -266,6 +266,11 @@ (logand result-bits #x7fffffff) (logand ,(expected-answer vector) #x7fffffff))))))))) +(defun make-expected-result-form (vector) + (if (eq (test vector) :eq) + (expected-answer vector) + t)) + (defun make-test-name (vector type) (intern (format nil "~@:(~A~)-~@:(~A~)-~@:(~A~).~D" (precision vector) @@ -298,7 +303,7 @@ `(make-double-float ,x)) (fun-args vector))))) ,(make-result-test-form vector))) - t) + ,(make-expected-result-form vector)) stream))) (defun emit-single-value-tests (vector stream) @@ -314,7 +319,7 @@ `(make-single-float ,x)) (fun-args vector))))) ,(make-result-test-form vector))) - t) + ,(make-expected-result-form vector)) stream))) (defun emit-double-exceptions-tests (vector stream) @@ -420,7 +425,7 @@ (dolist (type *float-types*) (pushnew (make-one-test-file fun type) *test-files* :test #'equal))) -(defvar *revision* "$Revision: 1.11 $") +(defvar *revision* "$Revision: 1.12 $") (defun format-date (stream arg colonp atp) (declare (ignore colonp atp))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieee754/ieee754-cmucl.lisp
by rtoy＠common-lisp.net 25 Aug '05

25 Aug '05

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ieee754 In directory common-lisp.net:/tmp/cvs-serv10146/ieee754 Modified Files: ieee754-cmucl.lisp Log Message: Add get-floating-point-modes. Date: Thu Aug 25 18:43:43 2005 Author: rtoy Index: ieeefp-tests/ieee754/ieee754-cmucl.lisp diff -u ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.3 ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.4 --- ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.3 Thu Jun 17 19:45:50 2004 +++ ieeefp-tests/ieee754/ieee754-cmucl.lisp Thu Aug 25 18:43:41 2005 @@ -30,6 +30,9 @@ precision)) (apply #'ext:set-floating-point-modes args)) +(defun get-floating-point-modes () + (ext:get-floating-point-modes)) + ;;; IEEE754 recommended functions

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieee754/ieee754-abcl.lisp
by pgraves＠common-lisp.net 25 Aug '05

25 Aug '05

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ieee754 In directory common-lisp.net:/tmp/cvs-serv4938 Modified Files: ieee754-abcl.lisp Log Message: Single float support, comparison operators. Date: Thu Aug 25 17:27:45 2005 Author: pgraves Index: ieeefp-tests/ieee754/ieee754-abcl.lisp diff -u ieeefp-tests/ieee754/ieee754-abcl.lisp:1.3 ieeefp-tests/ieee754/ieee754-abcl.lisp:1.4 --- ieeefp-tests/ieee754/ieee754-abcl.lisp:1.3 Tue Jun 15 21:58:31 2004 +++ ieeefp-tests/ieee754/ieee754-abcl.lisp Thu Aug 25 17:27:45 2005 @@ -1,27 +1,48 @@ (in-package "IEEE754-INTERNALS") -(defvar ieeefp-tests:*float-types* (list :double)) +(defvar ieeefp-tests:*float-types* (list :single :double)) (defvar ieeefp-tests:*rounding-modes* (list :nearest)) (defun make-single-float (x) - (error "Not supported.")) + (sys:make-single-float x)) (defun make-double-float (x) - (sys::make-double-float x)) + (sys:make-double-float x)) (defun single-float-bits (x) - (error "Not supported.")) + (declare (type single-float x)) + (ldb (byte 32 0) (sys:single-float-bits x))) (defun double-float-bits (x) (declare (type double-float x)) (ldb (byte 64 0) - (logior (ash (sys::double-float-high-bits x) 32) - (sys::double-float-low-bits x)))) + (logior (ash (sys:double-float-high-bits x) 32) + (sys:double-float-low-bits x)))) (defun set-floating-point-modes (&rest args &key traps accrued-exceptions current-exceptions rounding-mode precision) - (declare (ignore traps accrued-exceptions current-exceptions rounding-mode - precision)) + (declare (ignore args traps accrued-exceptions current-exceptions + rounding-mode precision)) ;; Not supported. ) + +(macrolet + ((def (x &body body) + `(defun ,x (x y) + (declare (type float x y)) + ,@body))) + (def = (cl:= x y)) + (def ?<> (not (= x y))) + (def > (cl:> x y)) + (def >= (cl:>= x y)) + (def < (cl:< x y)) + (def <= (cl:<= x y)) + (def ? (or (sys:float-nan-p x) (sys:float-nan-p y))) + (def <> (or (< x y) (> x y))) + (def <=> (or (< x y) (= x y) (> x y))) + (def ?> (or (? x y) (> x y))) + (def ?>= (or (? x y) (>= x y))) + (def ?< (or (? x y) (< x y))) + (def ?<= (or (? x y) (<= x y))) + (def ?= (or (? x y) (= x y))))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ucb-patches/ucblib/=d.input ieeefp-tests/ucb-patches/ucblib/=s.input
by Christophe Rhodes 02 Aug '04

02 Aug '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ucb-patches/ucblib In directory common-lisp.net:/tmp/cvs-serv24641/ucb-patches/ucblib Added Files: =d.input =s.input Log Message: Implement for sbcl comparison operators from Table 4 of ieee754. Work enough logic into ieeefp-tests to be able to test booleans as well as numerical results Add test vectors for ieee754:= Make sbcl test exceptions by default Date: Mon Aug 2 05:54:13 2004 Author: crhodes

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieeefp-tests.lisp ieeefp-tests/package.lisp
by Christophe Rhodes 02 Aug '04

02 Aug '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests In directory common-lisp.net:/tmp/cvs-serv24641 Modified Files: ieeefp-tests.lisp package.lisp Log Message: Implement for sbcl comparison operators from Table 4 of ieee754. Work enough logic into ieeefp-tests to be able to test booleans as well as numerical results Add test vectors for ieee754:= Make sbcl test exceptions by default Date: Mon Aug 2 05:54:12 2004 Author: crhodes Index: ieeefp-tests/ieeefp-tests.lisp diff -u ieeefp-tests/ieeefp-tests.lisp:1.10 ieeefp-tests/ieeefp-tests.lisp:1.11 --- ieeefp-tests/ieeefp-tests.lisp:1.10 Thu Jun 17 10:32:17 2004 +++ ieeefp-tests/ieeefp-tests.lisp Mon Aug 2 05:54:12 2004 @@ -37,9 +37,17 @@ (setf (fun-arity vector) 1)) ((add sub mul div pow atan2 hypot) (assert (= (length args-and-expected-answer) 3)) + (setf (fun-arity vector) 2)) + ((ieee754:=) + (assert (= (length args-and-expected-answer) 2)) (setf (fun-arity vector) 2))) - (setf (fun-args vector) (butlast args-and-expected-answer)) - (setf (expected-answer vector) (car (last args-and-expected-answer))) + (case (test vector) + ((:yea :nay) + (setf (fun-args vector) args-and-expected-answer + (expected-answer vector) nil)) + (t + (setf (fun-args vector) (butlast args-and-expected-answer)) + (setf (expected-answer vector) (car (last args-and-expected-answer))))) (setf (exceptions vector) (sort (exceptions vector) #'string<)) ;; FUN-NAME is currently partially overloaded with 2 meanings: 1. It ;; is the name of the test. 2. It is the name of the Lisp function @@ -106,7 +114,7 @@ (defun vector-pathname (function-name file-name) (let ((directory (case function-name - ((trunc) '(:relative "ucb-patches" "ucblib")) + ((trunc ieee754:=) '(:relative "ucb-patches" "ucblib")) (t '(:relative "ucb" "ucblib"))))) (merge-pathnames (make-pathname :directory directory @@ -183,12 +191,14 @@ stream list))) (defun make-result-test-form (vector) - `(if (complexp result) - t - ,(ecase (precision vector) - (:single (make-single-result-test-form vector)) - (:double (make-double-result-test-form vector)) - ))) + (case (test vector) + ((:yea) `(not (not result))) + ((:nay) `(not result)) + (t `(if (complexp result) + t + ,(ecase (precision vector) + (:single (make-single-result-test-form vector)) + (:double (make-double-result-test-form vector))))))) (defun make-double-result-test-form (vector) `(let ((result-bits (double-float-bits result))) @@ -406,11 +416,11 @@ (dolist (fun '(log sin cos tan sinh cosh tanh asin acos atan sqrt fabs floor ceil add sub mul div pow - atan2 log10 hypot trunc)) + atan2 log10 hypot trunc ieee754:=)) (dolist (type *float-types*) (pushnew (make-one-test-file fun type) *test-files* :test #'equal))) -(defvar *revision* "$Revision: 1.10 $") +(defvar *revision* "$Revision: 1.11 $") (defun format-date (stream arg colonp atp) (declare (ignore colonp atp)) Index: ieeefp-tests/package.lisp diff -u ieeefp-tests/package.lisp:1.4 ieeefp-tests/package.lisp:1.5 --- ieeefp-tests/package.lisp:1.4 Thu Jun 17 10:32:17 2004 +++ ieeefp-tests/package.lisp Mon Aug 2 05:54:12 2004 @@ -1,12 +1,21 @@ (defpackage "IEEE754" (:use "CL") + (:shadow "=" ">" "<" ">=" "<=") (:export "MAKE-SINGLE-FLOAT" "MAKE-DOUBLE-FLOAT" "SINGLE-FLOAT-BITS" "DOUBLE-FLOAT-BITS" - "GET-FLOATING-POINT-MODES" "SET-FLOATING-POINT-MODES")) + "GET-FLOATING-POINT-MODES" "SET-FLOATING-POINT-MODES" + + ;; IEEE754 Table 4 + "=" "?<>" ">" ">=" "<" "<=" "?" + "<>" + "<=>" "?>" "?>=" "?<" "?<=" "?=" + )) (defpackage "IEEE754-INTERNALS" + (:shadowing-import-from "IEEE754" "=" ">" "<" ">=" "<=") (:use "CL" "IEEE754")) (defpackage "IEEEFP-TESTS" + (:shadowing-import-from "CL" "=" ">" "<" ">=" "<=") (:use "CL" "IEEE754" "SPLIT-SEQUENCE") (:export "*FLOAT-TYPES*" "*ROUNDING-MODES*" "*TEST-EXCEPTIONS*"))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieee754/ieee754-sbcl.lisp
by Christophe Rhodes 02 Aug '04

02 Aug '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ieee754 In directory common-lisp.net:/tmp/cvs-serv24641/ieee754 Modified Files: ieee754-sbcl.lisp Log Message: Implement for sbcl comparison operators from Table 4 of ieee754. Work enough logic into ieeefp-tests to be able to test booleans as well as numerical results Add test vectors for ieee754:= Make sbcl test exceptions by default Date: Mon Aug 2 05:54:13 2004 Author: crhodes Index: ieeefp-tests/ieee754/ieee754-sbcl.lisp diff -u ieeefp-tests/ieee754/ieee754-sbcl.lisp:1.2 ieeefp-tests/ieee754/ieee754-sbcl.lisp:1.3 --- ieeefp-tests/ieee754/ieee754-sbcl.lisp:1.2 Thu Jun 17 10:32:17 2004 +++ ieeefp-tests/ieee754/ieee754-sbcl.lisp Mon Aug 2 05:54:12 2004 @@ -1,5 +1,7 @@ (in-package "IEEE754-INTERNALS") +(defvar ieeefp-tests:*test-exceptions* t) + (defun make-single-float (x) (declare (type (or (unsigned-byte 32) (signed-byte 32)) x)) (typecase x @@ -32,3 +34,23 @@ (defun get-floating-point-modes () (sb-int:get-floating-point-modes)) + +(macrolet + ((def (x &body body) + `(defun ,x (x y) + (declare (type float x y)) + ,@body))) + (def = (cl:= x y)) + (def ?<> (not (= x y))) + (def > (cl:> x y)) + (def >= (cl:>= x y)) + (def < (cl:< x y)) + (def <= (cl:<= x y)) + (def ? (or (sb-ext:float-nan-p x) (sb-ext:float-nan-p y))) + (def <> (or (< x y) (> x y))) + (def <=> (or (< x y) (= x y) (> x y))) + (def ?> (or (? x y) (> x y))) + (def ?>= (or (? x y) (>= x y))) + (def ?< (or (? x y) (< x y))) + (def ?<= (or (? x y) (<= x y))) + (def ?= (or (? x y) (= x y))))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieee754/ieee754-cmucl.lisp
by Raymond Toy 17 Jun '04

17 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ieee754 In directory common-lisp.net:/tmp/cvs-serv545 Modified Files: ieee754-cmucl.lisp Log Message: Misinterpreted what "after X in the direction of Y" meant. Fix it. It means adjust X to be closer to Y. (Noted by Christophe.) Date: Thu Jun 17 10:45:50 2004 Author: rtoy Index: ieeefp-tests/ieee754/ieee754-cmucl.lisp diff -u ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.2 ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.3 --- ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.2 Wed Jun 16 19:59:42 2004 +++ ieeefp-tests/ieee754/ieee754-cmucl.lisp Thu Jun 17 10:45:50 2004 @@ -81,7 +81,7 @@ (double-float -1022))))))) (defun nextafter (x y) - "The next double float after X in the direction of Y, with the + "The next float after X in the direction of Y, with the following exceptions: X = Y returns X, unchanged @@ -101,7 +101,7 @@ (t (multiple-value-bind (f e s) (integer-decode-float x) - (if (>= y 0) + (if (>= y x) (incf f) (decf f)) (* s (scale-float (float f x) e))))))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieeefp-tests.lisp ieeefp-tests/package.lisp
by Christophe Rhodes 17 Jun '04

17 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests In directory common-lisp.net:/tmp/cvs-serv23125 Modified Files: ieeefp-tests.lisp package.lisp Log Message: Add machinery for testing exceptions (disabled by default, enable using IEEEFP-TESTS:*TEST-EXCEPTIONS*). Include the original line in the generated test for slightly easier debugging. Date: Thu Jun 17 10:32:17 2004 Author: crhodes Index: ieeefp-tests/ieeefp-tests.lisp diff -u ieeefp-tests/ieeefp-tests.lisp:1.9 ieeefp-tests/ieeefp-tests.lisp:1.10 --- ieeefp-tests/ieeefp-tests.lisp:1.9 Wed Jun 16 03:35:54 2004 +++ ieeefp-tests/ieeefp-tests.lisp Thu Jun 17 10:32:17 2004 @@ -6,6 +6,8 @@ (defvar *rounding-modes* (list :nearest :zero :positive-infinity :negative-infinity)) +(defvar *test-exceptions* nil) + ;; So we can run log10 tests (defun log10 (x) (log x (float 10 x))) @@ -16,7 +18,8 @@ (abs (complex x y))) (defclass test-vector () - ((fun-name :initarg :fun-name :accessor fun-name) + ((line :initarg :line :accessor line) + (fun-name :initarg :fun-name :accessor fun-name) (lisp-fun-name :accessor lisp-fun-name) (fun-arity :accessor fun-arity) (precision :initarg :precision :accessor precision) @@ -144,7 +147,7 @@ ;; until we start testing exceptions. I think it ;; means that the following exception may or may not ;; be present. - (#\? (setf exceptions nil)))) + (#\? (push 'maybe exceptions)))) (setf args-and-expected-answer (loop for x on (nthcdr 4 split) by (ecase precision @@ -162,11 +165,12 @@ (ash (parse-integer (caddr x) :radix 16) 32) (parse-integer (cadddr x) :radix 16))))))) (push (make-instance 'test-vector + :line line :fun-name function-name :precision precision :rounding-mode rounding-mode :test test - :exceptions exceptions + :exceptions (nreverse exceptions) :args-and-expected-answer args-and-expected-answer) tests)))))))) @@ -268,116 +272,122 @@ (set-floating-point-modes :accrued-exceptions nil :current-exceptions nil)) +(defun get-accrued-exceptions () + (getf (get-floating-point-modes) :accrued-exceptions)) + (defun emit-double-value-tests (vector stream) - #| (when (eq (rounding-mode vector) :nearest) .. |# (when (member (rounding-mode vector) *rounding-modes*) (pprint - `(rt:deftest ,(make-test-name vector 'value #|'eval-value|#) + `(rt:deftest ,(make-test-name vector 'value) (progn + ,(line vector) (set-up-fpcw-state ,(rounding-mode vector)) (let ((result - #|(eval '|#(,(lisp-fun-name vector) - ,@(mapcar (lambda (x) - `(prog1 - (make-double-float ,x) - (clear-fpcw-exceptions))) - (fun-args vector)))))#|)|# + (,(lisp-fun-name vector) + ,@(mapcar (lambda (x) + `(make-double-float ,x)) + (fun-args vector))))) ,(make-result-test-form vector))) t) stream))) (defun emit-single-value-tests (vector stream) - #| (when (eq (rounding-mode vector) :nearest) .. |# (when (member (rounding-mode vector) *rounding-modes*) (pprint - `(rt:deftest ,(make-test-name vector 'value #|'eval-value|#) + `(rt:deftest ,(make-test-name vector 'value) (progn + ,(line vector) (set-up-fpcw-state ,(rounding-mode vector)) (let ((result - #|(eval '|#(,(lisp-fun-name vector) - ,@(mapcar (lambda (x) - `(prog1 - (make-single-float ,x) - (clear-fpcw-exceptions))) - (fun-args vector)))))#|)|# + (,(lisp-fun-name vector) + ,@(mapcar (lambda (x) + `(make-single-float ,x)) + (fun-args vector))))) ,(make-result-test-form vector))) t) - stream) - #+nil + stream))) + +(defun emit-double-exceptions-tests (vector stream) + (when (and (member (rounding-mode vector) *rounding-modes*) + *test-exceptions*) (pprint - `(rt:deftest ,(make-test-name vector 'compile-value) - (progn - ,(let ((arglist (mapcar (lambda (x) (declare (ignore x)) (gentemp)) - (fun-args vector)))) - `(let ((fn (compile nil '(lambda ,arglist - (,(lisp-fun-name vector) ,@arglist))))) - (set-up-fpcw-state ,(rounding-mode vector)) - (let ((result (funcall fn ,@(mapcar (lambda (x) - `(prog1 - (make-single-float ,x) - (clear-fpcw-exceptions))) - (fun-args vector))))) - ,(make-result-test-form vector))))) - t) - stream) - #+nil + (let* ((maybes (loop for x on (exceptions vector) + if (eq (car x) 'maybe) + collect (cadr x))) + (definites (sort + (set-difference (remove 'maybe (exceptions vector)) + maybes) + #'string<))) + `(rt:deftest ,(make-test-name vector 'exceptions) + (block nil + ,(line vector) + (set-up-fpcw-state ,(rounding-mode vector)) + (let ((answer + (,(lisp-fun-name vector) + ,@(mapcar (lambda (x) + `(make-double-float ,x)) + (fun-args vector)))) + (result + (sort + (remove-if-not + (lambda (x) + (member x + '(:invalid :underflow :overflow + :divide-by-zero :inexact))) + (set-difference (get-accrued-exceptions) + ',maybes)) + #'string<))) + (if (complexp answer) + ',definites + result))) + ,definites)) + stream))) + +(defun emit-single-exceptions-tests (vector stream) + (when (and (member (rounding-mode vector) *rounding-modes*) + *test-exceptions*) (pprint - `(rt:deftest ,(make-test-name vector 'compile-declared-value) - (progn - ,(let ((arglist (mapcar (lambda (x) (declare (ignore x)) (gentemp)) - (fun-args vector)))) - `(let ((fn (compile nil '(lambda ,arglist - (declare (type single-float ,@arglist)) - (,(lisp-fun-name vector) ,@arglist))))) - (set-up-fpcw-state ,(rounding-mode vector)) - (let ((result (funcall fn ,@(mapcar (lambda (x) - `(prog1 - (make-single-float ,x) - (clear-fpcw-exceptions))) - (fun-args vector))))) - ,(make-result-test-form vector))))) - t) + (let* ((maybes (loop for x on (exceptions vector) + if (eq (car x) 'maybe) + collect (cadr x))) + (definites (sort + (set-difference (remove 'maybe (exceptions vector)) + maybes) + #'string<))) + `(rt:deftest ,(make-test-name vector 'exceptions) + (block nil + ,(line vector) + (set-up-fpcw-state ,(rounding-mode vector)) + (let ((answer + (,(lisp-fun-name vector) + ,@(mapcar (lambda (x) + `(make-single-float ,x)) + (fun-args vector)))) + (result + (sort + (remove-if-not + (lambda (x) + (member x + '(:invalid :underflow :overflow + :divide-by-zero :inexact))) + (set-difference (get-accrued-exceptions) + ',maybes)) + #'string<))) + (if (complexp answer) + ',definites + result))) + ,definites)) stream))) (defmethod emit-tests-from-one-vector ((vector test-vector) stream) (let ((*print-case* :downcase)) (ecase (precision vector) (:single - #+nil - (pprint - `(rt:deftest ,(intern - (format nil "~@:(~A~)-~@:(~A~)-EVAL-EXCEPTIONS.~D" - (precision vector) - (lisp-fun-name vector) - *test-counter*)) - (progn - (set-floating-point-modes - :traps nil - :accrued-exceptions nil - :current-exceptions nil - :rounding-mode ,(rounding-mode vector)) - (let ((result - (eval '(,(lisp-fun-name vector) ,@(mapcar - (lambda (x) - `(prog1 - (make-single-float ,x) - (set-floating-point-modes - :accrued-exceptions nil - :current-exceptions nil))) - (fun-args vector)))))) - (if (complexp result) - ;; FIXME - ',(exceptions vector) - (sort - (intersection - (getf (sb-int:get-floating-point-modes) :accrued-exceptions) - '(:inexact :invalid :overflow :underflow :divide-by-zero)) - #'string<)))) - ,(exceptions vector)) - stream) - (emit-single-value-tests vector stream)) + (emit-single-value-tests vector stream) + (emit-single-exceptions-tests vector stream)) (:double - (emit-double-value-tests vector stream)) + (emit-double-value-tests vector stream) + (emit-double-exceptions-tests vector stream)) ))) (defun make-one-test-file (fun-name precision) @@ -400,7 +410,7 @@ (dolist (type *float-types*) (pushnew (make-one-test-file fun type) *test-files* :test #'equal))) -(defvar *revision* "$Revision: 1.9 $") +(defvar *revision* "$Revision: 1.10 $") (defun format-date (stream arg colonp atp) (declare (ignore colonp atp)) Index: ieeefp-tests/package.lisp diff -u ieeefp-tests/package.lisp:1.3 ieeefp-tests/package.lisp:1.4 --- ieeefp-tests/package.lisp:1.3 Tue Jun 15 06:55:07 2004 +++ ieeefp-tests/package.lisp Thu Jun 17 10:32:17 2004 @@ -2,11 +2,11 @@ (:use "CL") (:export "MAKE-SINGLE-FLOAT" "MAKE-DOUBLE-FLOAT" "SINGLE-FLOAT-BITS" "DOUBLE-FLOAT-BITS" - "SET-FLOATING-POINT-MODES")) + "GET-FLOATING-POINT-MODES" "SET-FLOATING-POINT-MODES")) (defpackage "IEEE754-INTERNALS" (:use "CL" "IEEE754")) (defpackage "IEEEFP-TESTS" (:use "CL" "IEEE754" "SPLIT-SEQUENCE") - (:export "*FLOAT-TYPES*" "*ROUNDING-MODES*")) + (:export "*FLOAT-TYPES*" "*ROUNDING-MODES*" "*TEST-EXCEPTIONS*"))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieee754/ieee754-sbcl.lisp
by Christophe Rhodes 17 Jun '04

17 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ieee754 In directory common-lisp.net:/tmp/cvs-serv23125/ieee754 Modified Files: ieee754-sbcl.lisp Log Message: Add machinery for testing exceptions (disabled by default, enable using IEEEFP-TESTS:*TEST-EXCEPTIONS*). Include the original line in the generated test for slightly easier debugging. Date: Thu Jun 17 10:32:17 2004 Author: crhodes Index: ieeefp-tests/ieee754/ieee754-sbcl.lisp diff -u ieeefp-tests/ieee754/ieee754-sbcl.lisp:1.1 ieeefp-tests/ieee754/ieee754-sbcl.lisp:1.2 --- ieeefp-tests/ieee754/ieee754-sbcl.lisp:1.1 Mon Jun 7 15:16:30 2004 +++ ieeefp-tests/ieee754/ieee754-sbcl.lisp Thu Jun 17 10:32:17 2004 @@ -30,3 +30,5 @@ precision)) (apply #'sb-int:set-floating-point-modes args)) +(defun get-floating-point-modes () + (sb-int:get-floating-point-modes))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieee754/ieee754-cmucl.lisp
by Raymond Toy 17 Jun '04

17 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ieee754 In directory common-lisp.net:/tmp/cvs-serv6045/ieee754 Modified Files: ieee754-cmucl.lisp Log Message: First pass at some of the IEEE754 recommended functions. Simple tests work, but still need to write a test-suite for these. Date: Wed Jun 16 19:59:42 2004 Author: rtoy Index: ieeefp-tests/ieee754/ieee754-cmucl.lisp diff -u ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.1 ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.2 --- ieeefp-tests/ieee754/ieee754-cmucl.lisp:1.1 Tue Jun 8 06:44:43 2004 +++ ieeefp-tests/ieee754/ieee754-cmucl.lisp Wed Jun 16 19:59:42 2004 @@ -30,3 +30,78 @@ precision)) (apply #'ext:set-floating-point-modes args)) + +;;; IEEE754 recommended functions + +(defun copysign (x y) + "Copy the sign of Y to X and return the result" + (float-sign y x)) + +(defun scalb (x n) + "Compute x*2^n, without roundoff" + (scale-float x n)) + +(defun finitep (x) + "Returns non-NIL if X is a finite value" + ;; What are we supposed to do when X is a NaN? + (not (ext:float-infinity-p x))) + +(defun nanp (x) + "Returns non-NIL if X is a NaN" + (ext:float-nan-p x)) + +(defun logb (x) + "Return the unbiased exponent of X, except for the following: + + NaN NaN + +infinity +infinity + 0 -infinity, signaling division by zero + + Also, 0 < scalb(x, -logb(x)) < 2, when x is positive and finite; it + is less than 1 only when x is denormalized." + (cond ((nanp x) + x) + ((not (finitep x)) + x) + ((zerop x) + ;; Need to signal division by zero, if enabled; otherwise, + ;; return -infinity. + (/ -1 x)) + (t + ;; Finite x. DECODE-FLOAT is basically what we want, except + ;; it's one too big. And it's also too small for + ;; denormalized numbers. We need to clip at the least + ;; exponent for normalized floats. + (multiple-value-bind (f e s) + (decode-float x) + (declare (ignore f s)) + (max (1- e) + (etypecase x + (single-float -126) + (double-float -1022))))))) + +(defun nextafter (x y) + "The next double float after X in the direction of Y, with the +following exceptions: + + X = Y returns X, unchanged + X or Y is quiet NaN returns the NaN + + Overflow is signaled if nextafter would overflow; underflow is +signaled if nextafter would underflow. In both cases, inexact +is signaled." + + ;; What are we supposed to do if x or y is a signaling NaN? + (cond ((= x y) + x) + ((nanp x) + x) + ((nanp y) + y) + (t + (multiple-value-bind (f e s) + (integer-decode-float x) + (if (>= y 0) + (incf f) + (decf f)) + (* s (scale-float (float f x) e))))))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ucb-patches/ucblib/truncd.input ieeefp-tests/ucb-patches/ucblib/truncs.input
by Christophe Rhodes 16 Jun '04

16 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ucb-patches/ucblib In directory common-lisp.net:/tmp/cvs-serv22284/ucb-patches/ucblib Added Files: truncd.input truncs.input Log Message: Add tests for trunc() Minor cleanups of asdf system description and of ieeefp-tests.lisp Date: Wed Jun 16 03:35:54 2004 Author: crhodes

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieeefp-tests.asd ieeefp-tests/ieeefp-tests.lisp
by Christophe Rhodes 16 Jun '04

16 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests In directory common-lisp.net:/tmp/cvs-serv22284 Modified Files: ieeefp-tests.asd ieeefp-tests.lisp Log Message: Add tests for trunc() Minor cleanups of asdf system description and of ieeefp-tests.lisp Date: Wed Jun 16 03:35:54 2004 Author: crhodes Index: ieeefp-tests/ieeefp-tests.asd diff -u ieeefp-tests/ieeefp-tests.asd:1.4 ieeefp-tests/ieeefp-tests.asd:1.5 --- ieeefp-tests/ieeefp-tests.asd:1.4 Tue Jun 15 06:55:07 2004 +++ ieeefp-tests/ieeefp-tests.asd Wed Jun 16 03:35:54 2004 @@ -18,7 +18,10 @@ (cl:defmethod asdf:perform ((o asdf:test-op) (c (cl:eql (asdf:find-system :ieeefp-tests)))) - (cl:mapcar #'cl:load - (cl:symbol-value (cl:intern "*TEST-FILES*" - (cl:find-package "IEEEFP-TESTS")))) + (cl:mapcar (cl:lambda (x) + (cl:format cl:*trace-output* "; loading ~S~%" (namestring x)) + (cl:load x :verbose nil)) + (cl:reverse + (cl:symbol-value (cl:intern "*TEST-FILES*" + (cl:find-package "IEEEFP-TESTS"))))) (cl:funcall (cl:intern "DO-TESTS" (cl:find-package "RT")))) Index: ieeefp-tests/ieeefp-tests.lisp diff -u ieeefp-tests/ieeefp-tests.lisp:1.8 ieeefp-tests/ieeefp-tests.lisp:1.9 --- ieeefp-tests/ieeefp-tests.lisp:1.8 Tue Jun 15 15:55:08 2004 +++ ieeefp-tests/ieeefp-tests.lisp Wed Jun 16 03:35:54 2004 @@ -29,7 +29,7 @@ (defmethod initialize-instance :after ((vector test-vector) &key args-and-expected-answer) (ecase (fun-name vector) - ((log exp sin cos tan sinh cosh tanh asin acos atan sqrt fabs floor ceil log10) + ((log exp sin cos tan sinh cosh tanh asin acos atan sqrt fabs floor ceil log10 trunc) (assert (= (length args-and-expected-answer) 2)) (setf (fun-arity vector) 1)) ((add sub mul div pow atan2 hypot) @@ -46,6 +46,7 @@ ((fabs) 'abs) ((floor) 'ffloor) ((ceil) 'fceiling) + ((trunc) 'ftruncate) ((add) '+) ((sub) '-) ((mul) '*) @@ -62,105 +63,59 @@ ((atan2) 'atan) (t (fun-name vector))))) -;; FIXME. This needs to be macroized or something so that each -;; constant is automatically added to *special-values* +(defparameter *special-values* nil) -(eval-when (:compile-toplevel :load-toplevel :execute) - ;; A bunch of constants for common IEEE values so we can read the - ;; tests more easily. - (defconstant +quiet-double-float-nan-mask+ - #X7FF8000000000000) - (defconstant +quiet-single-float-nan-mask+ - #X7FC00000) +(macrolet ((def (name bits) + `(progn + (defconstant ,name ,bits) + (push (cons ,bits ',name) *special-values*)))) + (def +quiet-double-float-nan-mask+ #X7FF8000000000000) + (def +quiet-single-float-nan-mask+ #X7FC00000) ;; NaN is supposed to ignore the sign, but the tests use both ;; positive and negative NaNs, so define them here. - (defconstant +trapping-positive-double-float-nan+ - #x7FF0000000000001) - (defconstant +trapping-negative-double-float-nan+ - #xFFF0000000000001) - (defconstant +single-float-positive-infinity+ - #x7F800000) - (defconstant +single-float-negative-infinity+ - #xFF800000) - (defconstant +double-float-positive-infinity+ - #x7FF0000000000000) - (defconstant +double-float-negative-infinity+ - #xFFF0000000000000) - (defconstant +most-positive-double-float+ - #X7FEFFFFFFFFFFFFF) - (defconstant +most-positive-single-float+ - #x7F7FFFFF) - (defconstant +least-positive-double-float+ - 1) - (defconstant +least-positive-single-float+ - 1) - (defconstant +least-negative-double-float+ - #x8000000000000001) - (defconstant +least-negative-single-float+ - #x80000001) - (defconstant +least-positive-normalized-double-float+ - #x10000000000000) - (defconstant +least-positive-normalized-single-float+ - #x800000) - (defconstant +least-negative-normalized-double-float+ - #x8010000000000000) - (defconstant +least-negative-normalized-single-float+ - #x80800000) - (defconstant +1d0+ - #x3FF0000000000000) - (defconstant +1f0+ - #x3F800000) - (defconstant +negative-0d0+ - #x8000000000000000) - (defconstant +negative-0f0+ - #x80000000) - ) - -;; An alist of integers and the corresponding symbol with that value. -(defparameter *special-values* - (mapcar #'(lambda (x) - `(,(symbol-value x) ,x)) - '(+quiet-double-float-nan-mask+ - +quiet-single-float-nan-mask+ - +trapping-positive-double-float-nan+ - +trapping-negative-double-float-nan+ - +single-float-positive-infinity+ - +single-float-negative-infinity+ - +double-float-positive-infinity+ - +double-float-negative-infinity+ - +most-positive-double-float+ - +most-positive-single-float+ - +least-positive-double-float+ - +least-positive-single-float+ - +least-negative-double-float+ - +least-negative-single-float+ - +least-positive-normalized-double-float+ - +least-positive-normalized-single-float+ - +least-negative-normalized-double-float+ - +least-negative-normalized-single-float+ - +1d0+ - +1f0+ - +negative-0d0+ - +negative-0f0+ - ))) + (def +trapping-positive-double-float-nan+ #x7FF0000000000001) + (def +trapping-negative-double-float-nan+ #xFFF0000000000001) + (def +single-float-positive-infinity+ #x7F800000) + (def +single-float-negative-infinity+ #xFF800000) + (def +double-float-positive-infinity+ #x7FF0000000000000) + (def +double-float-negative-infinity+ #xFFF0000000000000) + (def +most-positive-double-float+ #X7FEFFFFFFFFFFFFF) + (def +most-positive-single-float+ #x7F7FFFFF) + (def +least-positive-double-float+ 1) + (def +least-positive-single-float+ 1) + (def +least-negative-double-float+ #x8000000000000001) + (def +least-negative-single-float+ #x80000001) + (def +least-positive-normalized-double-float+ #x10000000000000) + (def +least-positive-normalized-single-float+ #x800000) + (def +least-negative-normalized-double-float+ #x8010000000000000) + (def +least-negative-normalized-single-float+ #x80800000) + (def +1d0+ #x3FF0000000000000) + (def +1f0+ #x3F800000) + (def +negative-0d0+ #x8000000000000000) + (def +negative-0f0+ #x80000000)) - (defun maybe-replace-special-value (x) ;; Look at x and replace it with named constants, if possible. (let ((value (assoc x *special-values* :test #'=))) (if value - (cadr value) + (cdr value) x))) +(defun vector-pathname (function-name file-name) + (let ((directory (case function-name + ((trunc) '(:relative "ucb-patches" "ucblib")) + (t '(:relative "ucb" "ucblib"))))) + (merge-pathnames + (make-pathname :directory directory + :name file-name + :type "input") + *load-truename*))) + (defun process-vector-file (function-name precision) (let* ((file-name (format nil "~(~A~)~(~A~)" function-name (char (symbol-name precision) 0))) (length (length file-name)) - (input-file (merge-pathnames - (make-pathname :directory '(:relative "ucb" "ucblib") - :name file-name - :type "input") - *load-truename*)) + (input-file (vector-pathname function-name file-name)) tests) (with-open-file (in input-file) @@ -429,7 +384,7 @@ (with-open-file (s (format nil "/tmp/~(~A~)-~(~A~).lisp" fun-name precision) :direction :output :if-exists :supersede) - (format t "; Creating ~S~%" (file-namestring s)) + (format *trace-output* "; creating ~S~%" (namestring s)) (format s "(in-package \"IEEEFP-TESTS\")~2%") (setf *test-counter* 0) (dolist (v (process-vector-file fun-name precision)) @@ -441,11 +396,11 @@ (dolist (fun '(log sin cos tan sinh cosh tanh asin acos atan sqrt fabs floor ceil add sub mul div pow - atan2 log10 hypot)) + atan2 log10 hypot trunc)) (dolist (type *float-types*) (pushnew (make-one-test-file fun type) *test-files* :test #'equal))) -(defvar *revision* "$Revision: 1.8 $") +(defvar *revision* "$Revision: 1.9 $") (defun format-date (stream arg colonp atp) (declare (ignore colonp atp))

1 0

[ieeefp-tests-cvs] CVS update: Directory change: ieeefp-tests/ucb-patches/ucblib
by Christophe Rhodes 16 Jun '04

16 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ucb-patches/ucblib In directory common-lisp.net:/tmp/cvs-serv10816/ucb-patches/ucblib Log Message: Directory /project/ieeefp-tests/cvsroot/ieeefp-tests/ucb-patches/ucblib added to the repository Date: Wed Jun 16 03:28:56 2004 Author: crhodes New directory ieeefp-tests/ucb-patches/ucblib added

1 0

[ieeefp-tests-cvs] CVS update: Directory change: ieeefp-tests/ucb-patches
by Christophe Rhodes 16 Jun '04

16 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ucb-patches In directory common-lisp.net:/tmp/cvs-serv10670/ucb-patches Log Message: Directory /project/ieeefp-tests/cvsroot/ieeefp-tests/ucb-patches added to the repository Date: Wed Jun 16 03:28:48 2004 Author: crhodes New directory ieeefp-tests/ucb-patches added

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieeefp-tests.lisp
by Raymond Toy 15 Jun '04

15 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests In directory common-lisp.net:/tmp/cvs-serv26648 Modified Files: ieeefp-tests.lisp Log Message: Oops. Fix typo. Date: Tue Jun 15 15:55:08 2004 Author: rtoy Index: ieeefp-tests/ieeefp-tests.lisp diff -u ieeefp-tests/ieeefp-tests.lisp:1.7 ieeefp-tests/ieeefp-tests.lisp:1.8 --- ieeefp-tests/ieeefp-tests.lisp:1.7 Tue Jun 15 15:51:19 2004 +++ ieeefp-tests/ieeefp-tests.lisp Tue Jun 15 15:55:08 2004 @@ -96,7 +96,7 @@ 1) (defconstant +least-negative-double-float+ #x8000000000000001) - (defconstant +least-positive-single-float+ + (defconstant +least-negative-single-float+ #x80000001) (defconstant +least-positive-normalized-double-float+ #x10000000000000) @@ -445,7 +445,7 @@ (dolist (type *float-types*) (pushnew (make-one-test-file fun type) *test-files* :test #'equal))) -(defvar *revision* "$Revision: 1.7 $") +(defvar *revision* "$Revision: 1.8 $") (defun format-date (stream arg colonp atp) (declare (ignore colonp atp))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieeefp-tests.lisp
by Raymond Toy 15 Jun '04

15 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests In directory common-lisp.net:/tmp/cvs-serv22606 Modified Files: ieeefp-tests.lisp Log Message: Add more constants. Date: Tue Jun 15 15:51:19 2004 Author: rtoy Index: ieeefp-tests/ieeefp-tests.lisp diff -u ieeefp-tests/ieeefp-tests.lisp:1.6 ieeefp-tests/ieeefp-tests.lisp:1.7 --- ieeefp-tests/ieeefp-tests.lisp:1.6 Tue Jun 15 15:23:29 2004 +++ ieeefp-tests/ieeefp-tests.lisp Tue Jun 15 15:51:19 2004 @@ -62,6 +62,9 @@ ((atan2) 'atan) (t (fun-name vector))))) +;; FIXME. This needs to be macroized or something so that each +;; constant is automatically added to *special-values* + (eval-when (:compile-toplevel :load-toplevel :execute) ;; A bunch of constants for common IEEE values so we can read the ;; tests more easily. @@ -69,7 +72,11 @@ #X7FF8000000000000) (defconstant +quiet-single-float-nan-mask+ #X7FC00000) - (defconstant +trapping-double-float-nan+ + ;; NaN is supposed to ignore the sign, but the tests use both + ;; positive and negative NaNs, so define them here. + (defconstant +trapping-positive-double-float-nan+ + #x7FF0000000000001) + (defconstant +trapping-negative-double-float-nan+ #xFFF0000000000001) (defconstant +single-float-positive-infinity+ #x7F800000) @@ -83,6 +90,22 @@ #X7FEFFFFFFFFFFFFF) (defconstant +most-positive-single-float+ #x7F7FFFFF) + (defconstant +least-positive-double-float+ + 1) + (defconstant +least-positive-single-float+ + 1) + (defconstant +least-negative-double-float+ + #x8000000000000001) + (defconstant +least-positive-single-float+ + #x80000001) + (defconstant +least-positive-normalized-double-float+ + #x10000000000000) + (defconstant +least-positive-normalized-single-float+ + #x800000) + (defconstant +least-negative-normalized-double-float+ + #x8010000000000000) + (defconstant +least-negative-normalized-single-float+ + #x80800000) (defconstant +1d0+ #x3FF0000000000000) (defconstant +1f0+ @@ -99,13 +122,22 @@ `(,(symbol-value x) ,x)) '(+quiet-double-float-nan-mask+ +quiet-single-float-nan-mask+ - +trapping-double-float-nan+ + +trapping-positive-double-float-nan+ + +trapping-negative-double-float-nan+ +single-float-positive-infinity+ +single-float-negative-infinity+ +double-float-positive-infinity+ +double-float-negative-infinity+ +most-positive-double-float+ +most-positive-single-float+ + +least-positive-double-float+ + +least-positive-single-float+ + +least-negative-double-float+ + +least-negative-single-float+ + +least-positive-normalized-double-float+ + +least-positive-normalized-single-float+ + +least-negative-normalized-double-float+ + +least-negative-normalized-single-float+ +1d0+ +1f0+ +negative-0d0+ @@ -397,6 +429,7 @@ (with-open-file (s (format nil "/tmp/~(~A~)-~(~A~).lisp" fun-name precision) :direction :output :if-exists :supersede) + (format t "; Creating ~S~%" (file-namestring s)) (format s "(in-package \"IEEEFP-TESTS\")~2%") (setf *test-counter* 0) (dolist (v (process-vector-file fun-name precision)) @@ -412,7 +445,7 @@ (dolist (type *float-types*) (pushnew (make-one-test-file fun type) *test-files* :test #'equal))) -(defvar *revision* "$Revision: 1.6 $") +(defvar *revision* "$Revision: 1.7 $") (defun format-date (stream arg colonp atp) (declare (ignore colonp atp))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieeefp-tests.lisp
by Raymond Toy 15 Jun '04

15 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests In directory common-lisp.net:/tmp/cvs-serv19142 Modified Files: ieeefp-tests.lisp Log Message: o Add support for hypot. This is used to test Lisp's ABS function on complex values. o Clean up some comments, remove FIXME about atan2, log10, hypot. Date: Tue Jun 15 15:23:29 2004 Author: rtoy Index: ieeefp-tests/ieeefp-tests.lisp diff -u ieeefp-tests/ieeefp-tests.lisp:1.5 ieeefp-tests/ieeefp-tests.lisp:1.6 --- ieeefp-tests/ieeefp-tests.lisp:1.5 Tue Jun 15 15:03:48 2004 +++ ieeefp-tests/ieeefp-tests.lisp Tue Jun 15 15:23:29 2004 @@ -10,6 +10,11 @@ (defun log10 (x) (log x (float 10 x))) +;; So we can run hypot tests and make it test Lisp's ABS function on +;; complex values. +(defun hypot (x y) + (abs (complex x y))) + (defclass test-vector () ((fun-name :initarg :fun-name :accessor fun-name) (lisp-fun-name :accessor lisp-fun-name) @@ -24,18 +29,18 @@ (defmethod initialize-instance :after ((vector test-vector) &key args-and-expected-answer) (ecase (fun-name vector) - ;; FIXME: atan comes in two versions; log10 exists; then there's - ;; hypot() and cabs() which appear not to have equivalents in CL. - ;; (Could use them to test ABS on complexes, though) ((log exp sin cos tan sinh cosh tanh asin acos atan sqrt fabs floor ceil log10) (assert (= (length args-and-expected-answer) 2)) (setf (fun-arity vector) 1)) - ((add sub mul div pow atan2) + ((add sub mul div pow atan2 hypot) (assert (= (length args-and-expected-answer) 3)) (setf (fun-arity vector) 2))) (setf (fun-args vector) (butlast args-and-expected-answer)) (setf (expected-answer vector) (car (last args-and-expected-answer))) (setf (exceptions vector) (sort (exceptions vector) #'string<)) + ;; FUN-NAME is currently partially overloaded with 2 meanings: 1. It + ;; is the name of the test. 2. It is the name of the Lisp function + ;; to use. (setf (fun-name vector) (case (fun-name vector) ((fabs) 'abs) @@ -54,14 +59,6 @@ ;; be atan, overwriting the tests for the single arg atan. (setf (lisp-fun-name vector) (case (fun-name vector) - ((fabs) 'abs) - ((floor) 'ffloor) - ((ceil) 'fceiling) - ((add) '+) - ((sub) '-) - ((mul) '*) - ((div) '/) - ((pow) 'expt) ((atan2) 'atan) (t (fun-name vector))))) @@ -411,11 +408,11 @@ (dolist (fun '(log sin cos tan sinh cosh tanh asin acos atan sqrt fabs floor ceil add sub mul div pow - atan2 log10)) + atan2 log10 hypot)) (dolist (type *float-types*) (pushnew (make-one-test-file fun type) *test-files* :test #'equal))) -(defvar *revision* "$Revision: 1.5 $") +(defvar *revision* "$Revision: 1.6 $") (defun format-date (stream arg colonp atp) (declare (ignore colonp atp))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieeefp-tests.lisp
by Raymond Toy 15 Jun '04

15 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests In directory common-lisp.net:/tmp/cvs-serv2301 Modified Files: ieeefp-tests.lisp Log Message: Add support for atan2 and log10 tests: o By adding an extra slot to TEST-VECTOR to hold the lisp function name we need to use. (Only really needed for atan2.) o By adding a log10 function for us to call. Date: Tue Jun 15 15:03:48 2004 Author: rtoy Index: ieeefp-tests/ieeefp-tests.lisp diff -u ieeefp-tests/ieeefp-tests.lisp:1.4 ieeefp-tests/ieeefp-tests.lisp:1.5 --- ieeefp-tests/ieeefp-tests.lisp:1.4 Wed Jun 9 09:05:17 2004 +++ ieeefp-tests/ieeefp-tests.lisp Tue Jun 15 15:03:48 2004 @@ -6,8 +6,13 @@ (defvar *rounding-modes* (list :nearest :zero :positive-infinity :negative-infinity)) +;; So we can run log10 tests +(defun log10 (x) + (log x (float 10 x))) + (defclass test-vector () ((fun-name :initarg :fun-name :accessor fun-name) + (lisp-fun-name :accessor lisp-fun-name) (fun-arity :accessor fun-arity) (precision :initarg :precision :accessor precision) (rounding-mode :initarg :rounding-mode :accessor rounding-mode) @@ -22,10 +27,10 @@ ;; FIXME: atan comes in two versions; log10 exists; then there's ;; hypot() and cabs() which appear not to have equivalents in CL. ;; (Could use them to test ABS on complexes, though) - ((log exp sin cos tan sinh cosh tanh asin acos atan sqrt fabs floor ceil) + ((log exp sin cos tan sinh cosh tanh asin acos atan sqrt fabs floor ceil log10) (assert (= (length args-and-expected-answer) 2)) (setf (fun-arity vector) 1)) - ((add sub mul div pow) + ((add sub mul div pow atan2) (assert (= (length args-and-expected-answer) 3)) (setf (fun-arity vector) 2))) (setf (fun-args vector) (butlast args-and-expected-answer)) @@ -41,6 +46,23 @@ ((mul) '*) ((div) '/) ((pow) 'expt) + (t (fun-name vector)))) + ;; Figure out the Lisp function we need to call to test. Mostly + ;; redundant, except for the atan2 tests. Can't use a fun-name of + ;; atan2 because there's no atan2 Lisp function. And can't change + ;; fun-name from atan2 to atan because then all the test names will + ;; be atan, overwriting the tests for the single arg atan. + (setf (lisp-fun-name vector) + (case (fun-name vector) + ((fabs) 'abs) + ((floor) 'ffloor) + ((ceil) 'fceiling) + ((add) '+) + ((sub) '-) + ((mul) '*) + ((div) '/) + ((pow) 'expt) + ((atan2) 'atan) (t (fun-name vector))))) (eval-when (:compile-toplevel :load-toplevel :execute) @@ -270,7 +292,7 @@ (progn (set-up-fpcw-state ,(rounding-mode vector)) (let ((result - #|(eval '|#(,(fun-name vector) + #|(eval '|#(,(lisp-fun-name vector) ,@(mapcar (lambda (x) `(prog1 (make-double-float ,x) @@ -288,7 +310,7 @@ (progn (set-up-fpcw-state ,(rounding-mode vector)) (let ((result - #|(eval '|#(,(fun-name vector) + #|(eval '|#(,(lisp-fun-name vector) ,@(mapcar (lambda (x) `(prog1 (make-single-float ,x) @@ -304,7 +326,7 @@ ,(let ((arglist (mapcar (lambda (x) (declare (ignore x)) (gentemp)) (fun-args vector)))) `(let ((fn (compile nil '(lambda ,arglist - (,(fun-name vector) ,@arglist))))) + (,(lisp-fun-name vector) ,@arglist))))) (set-up-fpcw-state ,(rounding-mode vector)) (let ((result (funcall fn ,@(mapcar (lambda (x) `(prog1 @@ -322,7 +344,7 @@ (fun-args vector)))) `(let ((fn (compile nil '(lambda ,arglist (declare (type single-float ,@arglist)) - (,(fun-name vector) ,@arglist))))) + (,(lisp-fun-name vector) ,@arglist))))) (set-up-fpcw-state ,(rounding-mode vector)) (let ((result (funcall fn ,@(mapcar (lambda (x) `(prog1 @@ -342,7 +364,7 @@ `(rt:deftest ,(intern (format nil "~@:(~A~)-~@:(~A~)-EVAL-EXCEPTIONS.~D" (precision vector) - (fun-name vector) + (lisp-fun-name vector) *test-counter*)) (progn (set-floating-point-modes @@ -351,7 +373,7 @@ :current-exceptions nil :rounding-mode ,(rounding-mode vector)) (let ((result - (eval '(,(fun-name vector) ,@(mapcar + (eval '(,(lisp-fun-name vector) ,@(mapcar (lambda (x) `(prog1 (make-single-float ,x) @@ -388,11 +410,12 @@ (defparameter *test-files* nil) (dolist (fun '(log sin cos tan sinh cosh tanh asin acos - atan sqrt fabs floor ceil add sub mul div pow)) + atan sqrt fabs floor ceil add sub mul div pow + atan2 log10)) (dolist (type *float-types*) (pushnew (make-one-test-file fun type) *test-files* :test #'equal))) -(defvar *revision* "$Revision: 1.4 $") +(defvar *revision* "$Revision: 1.5 $") (defun format-date (stream arg colonp atp) (declare (ignore colonp atp))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieee754/ieee754-abcl.lisp
by Peter Graves 15 Jun '04

15 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ieee754 In directory common-lisp.net:/tmp/cvs-serv17129 Modified Files: ieee754-abcl.lisp Log Message: The package name is "IEEEFP-TESTS", not "IEEE754-TESTS". Date: Tue Jun 15 12:58:31 2004 Author: pgraves Index: ieeefp-tests/ieee754/ieee754-abcl.lisp diff -u ieeefp-tests/ieee754/ieee754-abcl.lisp:1.2 ieeefp-tests/ieee754/ieee754-abcl.lisp:1.3 --- ieeefp-tests/ieee754/ieee754-abcl.lisp:1.2 Tue Jun 15 06:55:08 2004 +++ ieeefp-tests/ieee754/ieee754-abcl.lisp Tue Jun 15 12:58:31 2004 @@ -1,8 +1,8 @@ (in-package "IEEE754-INTERNALS") -(defvar ieee754-tests:*float-types* (list :double)) +(defvar ieeefp-tests:*float-types* (list :double)) -(defvar ieee754-tests:*rounding-modes* (list :nearest)) +(defvar ieeefp-tests:*rounding-modes* (list :nearest)) (defun make-single-float (x) (error "Not supported."))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieee754/ieee754-abcl.lisp
by Christophe Rhodes 15 Jun '04

15 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests/ieee754 In directory common-lisp.net:/tmp/cvs-serv24446/ieee754 Modified Files: ieee754-abcl.lisp Log Message: Move *ROUNDING-MODES* and *FLOAT-TYPES* to the IEEEFP-TESTS package. Make the asdf system description actually do what it's meant to. Date: Tue Jun 15 06:55:08 2004 Author: crhodes Index: ieeefp-tests/ieee754/ieee754-abcl.lisp diff -u ieeefp-tests/ieee754/ieee754-abcl.lisp:1.1 ieeefp-tests/ieee754/ieee754-abcl.lisp:1.2 --- ieeefp-tests/ieee754/ieee754-abcl.lisp:1.1 Wed Jun 9 08:51:02 2004 +++ ieeefp-tests/ieee754/ieee754-abcl.lisp Tue Jun 15 06:55:08 2004 @@ -1,8 +1,8 @@ (in-package "IEEE754-INTERNALS") -(defvar *float-types* (list :double)) +(defvar ieee754-tests:*float-types* (list :double)) -(defvar *rounding-modes* (list :nearest)) +(defvar ieee754-tests:*rounding-modes* (list :nearest)) (defun make-single-float (x) (error "Not supported."))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieeefp-tests.asd ieeefp-tests/package.lisp
by Christophe Rhodes 15 Jun '04

15 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests In directory common-lisp.net:/tmp/cvs-serv24446 Modified Files: ieeefp-tests.asd package.lisp Log Message: Move *ROUNDING-MODES* and *FLOAT-TYPES* to the IEEEFP-TESTS package. Make the asdf system description actually do what it's meant to. Date: Tue Jun 15 06:55:07 2004 Author: crhodes Index: ieeefp-tests/ieeefp-tests.asd diff -u ieeefp-tests/ieeefp-tests.asd:1.3 ieeefp-tests/ieeefp-tests.asd:1.4 --- ieeefp-tests/ieeefp-tests.asd:1.3 Wed Jun 9 08:51:44 2004 +++ ieeefp-tests/ieeefp-tests.asd Tue Jun 15 06:55:07 2004 @@ -5,14 +5,20 @@ :depends-on ("package") :components ((:file "ieee754-sbcl" - :in-order-to ((compile-op (feature :sbcl)))) + :in-order-to ((asdf:compile-op + (asdf:feature :sbcl)))) (:file "ieee754-cmucl" - :in-order-to ((compile-op (feature :cmu)))) + :in-order-to ((asdf:compile-op + (asdf:feature :cmu)))) (:file "ieee754-abcl" - :in-order-to ((compile-op (feature :abcl))))) + :in-order-to ((asdf:compile-op + (asdf:feature :abcl))))) :if-component-dep-fails :ignore) (:file "ieeefp-tests" :depends-on ("ieee754")))) -(defmethod asdf:perform ((o asdf:test-op) (c (eql (asdf:find-system :ieeefp-tests)))) - (mapcar #'load (symbol-value (intern "*TEST-FILES*" (find-package "IEEEFP-TESTS")))) - (funcall (intern "DO-TESTS" (find-package "RT")))) +(cl:defmethod asdf:perform ((o asdf:test-op) + (c (cl:eql (asdf:find-system :ieeefp-tests)))) + (cl:mapcar #'cl:load + (cl:symbol-value (cl:intern "*TEST-FILES*" + (cl:find-package "IEEEFP-TESTS")))) + (cl:funcall (cl:intern "DO-TESTS" (cl:find-package "RT")))) Index: ieeefp-tests/package.lisp diff -u ieeefp-tests/package.lisp:1.2 ieeefp-tests/package.lisp:1.3 --- ieeefp-tests/package.lisp:1.2 Wed Jun 9 08:53:42 2004 +++ ieeefp-tests/package.lisp Tue Jun 15 06:55:07 2004 @@ -1,7 +1,6 @@ (defpackage "IEEE754" (:use "CL") - (:export "*FLOAT-TYPES*" "*ROUNDING-MODES*" - "MAKE-SINGLE-FLOAT" "MAKE-DOUBLE-FLOAT" + (:export "MAKE-SINGLE-FLOAT" "MAKE-DOUBLE-FLOAT" "SINGLE-FLOAT-BITS" "DOUBLE-FLOAT-BITS" "SET-FLOATING-POINT-MODES")) @@ -9,4 +8,5 @@ (:use "CL" "IEEE754")) (defpackage "IEEEFP-TESTS" - (:use "CL" "IEEE754" "SPLIT-SEQUENCE")) \ No newline at end of file + (:use "CL" "IEEE754" "SPLIT-SEQUENCE") + (:export "*FLOAT-TYPES*" "*ROUNDING-MODES*"))

1 0

[ieeefp-tests-cvs] CVS update: ieeefp-tests/ieeefp-tests.lisp
by Peter Graves 09 Jun '04

09 Jun '04

Update of /project/ieeefp-tests/cvsroot/ieeefp-tests In directory common-lisp.net:/tmp/cvs-serv27365 Modified Files: ieeefp-tests.lisp Log Message: Added *FLOAT-TYPES* and *ROUNDING-MODES* to support customization of float types and rounding modes supported by an implementation. Date: Wed Jun 9 09:05:17 2004 Author: pgraves Index: ieeefp-tests/ieeefp-tests.lisp diff -u ieeefp-tests/ieeefp-tests.lisp:1.3 ieeefp-tests/ieeefp-tests.lisp:1.4 --- ieeefp-tests/ieeefp-tests.lisp:1.3 Wed Jun 9 06:06:17 2004 +++ ieeefp-tests/ieeefp-tests.lisp Wed Jun 9 09:05:17 2004 @@ -1,5 +1,11 @@ (in-package "IEEEFP-TESTS") +(defvar *float-types* + (list :single :double)) + +(defvar *rounding-modes* + (list :nearest :zero :positive-infinity :negative-infinity)) + (defclass test-vector () ((fun-name :initarg :fun-name :accessor fun-name) (fun-arity :accessor fun-arity) @@ -9,7 +15,7 @@ (exceptions :initarg :exceptions :accessor exceptions) (fun-args :accessor fun-args) (expected-answer :accessor expected-answer))) - + (defmethod initialize-instance :after ((vector test-vector) &key args-and-expected-answer) (ecase (fun-name vector) @@ -105,7 +111,7 @@ :type "input") *load-truename*)) tests) - + (with-open-file (in input-file) (do ((line (read-line in nil nil) (read-line in nil nil))) ((null line) (nreverse tests)) @@ -258,72 +264,74 @@ (defun emit-double-value-tests (vector stream) #| (when (eq (rounding-mode vector) :nearest) .. |# - (pprint - `(rt:deftest ,(make-test-name vector 'value #|'eval-value|#) - (progn - (set-up-fpcw-state ,(rounding-mode vector)) - (let ((result - #|(eval '|#(,(fun-name vector) - ,@(mapcar (lambda (x) - `(prog1 - (make-double-float ,x) - (clear-fpcw-exceptions))) - (fun-args vector)))))#|)|# - ,(make-result-test-form vector))) - t) - stream)) + (when (member (rounding-mode vector) *rounding-modes*) + (pprint + `(rt:deftest ,(make-test-name vector 'value #|'eval-value|#) + (progn + (set-up-fpcw-state ,(rounding-mode vector)) + (let ((result + #|(eval '|#(,(fun-name vector) + ,@(mapcar (lambda (x) + `(prog1 + (make-double-float ,x) + (clear-fpcw-exceptions))) + (fun-args vector)))))#|)|# + ,(make-result-test-form vector))) + t) + stream))) (defun emit-single-value-tests (vector stream) #| (when (eq (rounding-mode vector) :nearest) .. |# - (pprint - `(rt:deftest ,(make-test-name vector 'value #|'eval-value|#) - (progn - (set-up-fpcw-state ,(rounding-mode vector)) - (let ((result - #|(eval '|#(,(fun-name vector) - ,@(mapcar (lambda (x) - `(prog1 - (make-single-float ,x) - (clear-fpcw-exceptions))) - (fun-args vector)))))#|)|# - ,(make-result-test-form vector))) - t) - stream) - #+nil - (pprint - `(rt:deftest ,(make-test-name vector 'compile-value) - (progn - ,(let ((arglist (mapcar (lambda (x) (declare (ignore x)) (gentemp)) - (fun-args vector)))) - `(let ((fn (compile nil '(lambda ,arglist - (,(fun-name vector) ,@arglist))))) - (set-up-fpcw-state ,(rounding-mode vector)) - (let ((result (funcall fn ,@(mapcar (lambda (x) - `(prog1 - (make-single-float ,x) - (clear-fpcw-exceptions))) - (fun-args vector))))) - ,(make-result-test-form vector))))) - t) - stream) - #+nil - (pprint - `(rt:deftest ,(make-test-name vector 'compile-declared-value) - (progn - ,(let ((arglist (mapcar (lambda (x) (declare (ignore x)) (gentemp)) - (fun-args vector)))) - `(let ((fn (compile nil '(lambda ,arglist - (declare (type single-float ,@arglist)) - (,(fun-name vector) ,@arglist))))) - (set-up-fpcw-state ,(rounding-mode vector)) - (let ((result (funcall fn ,@(mapcar (lambda (x) - `(prog1 - (make-single-float ,x) - (clear-fpcw-exceptions))) - (fun-args vector))))) - ,(make-result-test-form vector))))) - t) - stream)) + (when (member (rounding-mode vector) *rounding-modes*) + (pprint + `(rt:deftest ,(make-test-name vector 'value #|'eval-value|#) + (progn + (set-up-fpcw-state ,(rounding-mode vector)) + (let ((result + #|(eval '|#(,(fun-name vector) + ,@(mapcar (lambda (x) + `(prog1 + (make-single-float ,x) + (clear-fpcw-exceptions))) + (fun-args vector)))))#|)|# + ,(make-result-test-form vector))) + t) + stream) + #+nil + (pprint + `(rt:deftest ,(make-test-name vector 'compile-value) + (progn + ,(let ((arglist (mapcar (lambda (x) (declare (ignore x)) (gentemp)) + (fun-args vector)))) + `(let ((fn (compile nil '(lambda ,arglist + (,(fun-name vector) ,@arglist))))) + (set-up-fpcw-state ,(rounding-mode vector)) + (let ((result (funcall fn ,@(mapcar (lambda (x) + `(prog1 + (make-single-float ,x) + (clear-fpcw-exceptions))) + (fun-args vector))))) + ,(make-result-test-form vector))))) + t) + stream) + #+nil + (pprint + `(rt:deftest ,(make-test-name vector 'compile-declared-value) + (progn + ,(let ((arglist (mapcar (lambda (x) (declare (ignore x)) (gentemp)) + (fun-args vector)))) + `(let ((fn (compile nil '(lambda ,arglist + (declare (type single-float ,@arglist)) + (,(fun-name vector) ,@arglist))))) + (set-up-fpcw-state ,(rounding-mode vector)) + (let ((result (funcall fn ,@(mapcar (lambda (x) + `(prog1 + (make-single-float ,x) + (clear-fpcw-exceptions))) + (fun-args vector))))) + ,(make-result-test-form vector))))) + t) + stream))) (defmethod emit-tests-from-one-vector ((vector test-vector) stream) (let ((*print-case* :downcase)) @@ -381,10 +389,10 @@ (dolist (fun '(log sin cos tan sinh cosh tanh asin acos atan sqrt fabs floor ceil add sub mul div pow)) - (pushnew (make-one-test-file fun :single) *test-files* :test #'equal) - (pushnew (make-one-test-file fun :double) *test-files* :test #'equal)) + (dolist (type *float-types*) + (pushnew (make-one-test-file fun type) *test-files* :test #'equal))) -(defvar *revision* "$Revision: 1.3 $") +(defvar *revision* "$Revision: 1.4 $") (defun format-date (stream arg colonp atp) (declare (ignore colonp atp))

1 0