[Bug 43959] New: webservices/reader tests fail on arm
https://bugs.winehq.org/show_bug.cgi?id=43959 Bug ID: 43959 Summary: webservices/reader tests fail on arm Product: Wine Version: 2.19 Hardware: arm OS: Linux Status: NEW Keywords: download, testcase Severity: normal Priority: P2 Component: -unknown Assignee: wine-bugs(a)winehq.org Reporter: austinenglish(a)gmail.com Distribution: Debian fixme:webservices:set_fpword not implemented reader.c:3733: Test failed: 0: got 80004001 fixme:webservices:set_fpword not implemented reader.c:3733: Test failed: 1: got 80004001 fixme:webservices:set_fpword not implemented reader.c:3733: Test failed: 2: got 80004001 Pretty obvious why: https://source.winehq.org/git/wine.git/blob/039d267b0925273197a9edcf7664c4a4... 3599 BOOL set_fpword( unsigned short new, unsigned short *old ) 3600 { 3601 #if defined(__GNUC__) && (defined(__i386__) || defined(__x86_64__)) 3602 unsigned short fpword; 3603 3604 __asm__ __volatile__( "fstcw %0" : "=m" (fpword) ); 3605 *old = fpword; 3606 fpword = new; 3607 __asm__ __volatile__( "fldcw %0" : : "m" (fpword) ); 3608 return TRUE; 3609 #else 3610 FIXME( "not implemented\n" ); 3611 return FALSE; 3612 #endif 3613 } -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 André H. <nerv(a)dawncrow.de> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |nerv(a)dawncrow.de --- Comment #1 from André H. <nerv(a)dawncrow.de> --- confirming: ... fixme:webservices:set_fpword not implemented reader.c:6263: Test failed: 51: got 80004001 reader.c:6388: Test marked todo: got 3 0008:reader: 3336 tests executed (46 marked as todo, 98 failures), 0 skipped. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 André H. <nerv(a)dawncrow.de> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |alexhenrie24(a)gmail.com --- Comment #2 from André H. <nerv(a)dawncrow.de> --- https://source.winehq.org/patches/data/138483 https://source.winehq.org/patches/data/138484 reader and writer tests don't get down to zero, but that seems to be precision related Adding Alex, as he is doing some precision related work for win64 tests atm it seems reader.c:3734: Test failed: 55: got 7ff0000000000000 reader.c:3734: Test failed: 57: got 0 writer.c:388: Test failed: 2280: got 11 expected 10 writer.c:390: Test failed: 2280: got <t>:E-3</t> expected <t>1E-2</t> writer.c:388: Test failed: 2280: got 12 expected 11 writer.c:390: Test failed: 2280: got <t>-:E-3</t> expected <t>-1E-2</t> writer.c:390: Test failed: 2280: got <t>1.7976931348623156E+308</t> expected <t>1.7976931348623157E+308</t> writer.c:390: Test failed: 2280: got <t>-1.7976931348623156E+308</t> expected <t>-1.7976931348623157E+308</t> -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 --- Comment #3 from Hans Leidekker <hans(a)meelstraat.net> --- (In reply to André H. from comment #2)
https://source.winehq.org/patches/data/138483 https://source.winehq.org/patches/data/138484
reader and writer tests don't get down to zero, but that seems to be precision related
Adding Alex, as he is doing some precision related work for win64 tests atm it seems
reader.c:3734: Test failed: 55: got 7ff0000000000000 reader.c:3734: Test failed: 57: got 0
writer.c:388: Test failed: 2280: got 11 expected 10 writer.c:390: Test failed: 2280: got <t>:E-3</t> expected <t>1E-2</t> writer.c:388: Test failed: 2280: got 12 expected 11 writer.c:390: Test failed: 2280: got <t>-:E-3</t> expected <t>-1E-2</t> writer.c:390: Test failed: 2280: got <t>1.7976931348623156E+308</t> expected <t>1.7976931348623157E+308</t> writer.c:390: Test failed: 2280: got <t>-1.7976931348623156E+308</t> expected <t>-1.7976931348623157E+308</t>
Do the tests pass on ARM Windows? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 --- Comment #4 from Alex Henrie <alexhenrie24(a)gmail.com> --- This does not look like a precision problem: writer.c:390: Test failed: 2280: got <t>:E-3</t> expected <t>1E-2</t> ':' == '9' + 1, so it looks like an off-by-one error. We should fix this first. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 --- Comment #5 from Hans Leidekker <hans(a)meelstraat.net> --- The code depends on the extra precision offered by the long double type on x86. It's an 80-bit quantity on x86 and 64-bit on ARM, AFAICT. Before we make changes to the code I'd like to confirm that the current tests reflect ARM Windows behavior. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 --- Comment #6 from Austin English <austinenglish(a)gmail.com> --- (In reply to Hans Leidekker from comment #5)
Before we make changes to the code I'd like to confirm that the current tests reflect ARM Windows behavior.
I just ordered a used Surface RT on ebay, so hopefully I can get more info on this in the coming days. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 --- Comment #7 from Austin English <austinenglish(a)gmail.com> --- Created attachment 67185 --> https://bugs.winehq.org/attachment.cgi?id=67185 surface rt output -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 --- Comment #8 from Austin English <austinenglish(a)gmail.com> --- (In reply to Hans Leidekker from comment #5)
The code depends on the extra precision offered by the long double type on x86. It's an 80-bit quantity on x86 and 64-bit on ARM, AFAICT.
Before we make changes to the code I'd like to confirm that the current tests reflect ARM Windows behavior.
It passes on windows arm (wasn't really worth attaching, oops): 0050:reader: 3624 tests executed (0 marked as todo, 0 failures), 0 skipped. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 --- Comment #9 from Hans Leidekker <hans(a)meelstraat.net> --- (In reply to Austin English from comment #8)
(In reply to Hans Leidekker from comment #5)
The code depends on the extra precision offered by the long double type on x86. It's an 80-bit quantity on x86 and 64-bit on ARM, AFAICT.
Before we make changes to the code I'd like to confirm that the current tests reflect ARM Windows behavior.
It passes on windows arm (wasn't really worth attaching, oops): 0050:reader: 3624 tests executed (0 marked as todo, 0 failures), 0 skipped.
There were changes in this code, does it still fail on Wine? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 --- Comment #10 from Austin English <austinenglish(a)gmail.com> --- (In reply to Hans Leidekker from comment #9)
(In reply to Austin English from comment #8)
(In reply to Hans Leidekker from comment #5)
The code depends on the extra precision offered by the long double type on x86. It's an 80-bit quantity on x86 and 64-bit on ARM, AFAICT.
Before we make changes to the code I'd like to confirm that the current tests reflect ARM Windows behavior.
It passes on windows arm (wasn't really worth attaching, oops): 0050:reader: 3624 tests executed (0 marked as todo, 0 failures), 0 skipped.
There were changes in this code, does it still fail on Wine?
reader.c now passes in wine-5.8-230-g3bb824f988. Currently bisected the fix. writer.c still fails: austin(a)PrawnOS:~/wine-git/dlls/webservices/tests$ WINEDEBUG=-all make writer.ok ../../../tools/runtest -q -P wine -T ../../.. -M webservices.dll -p webservices_test.exe.so writer && touch writer.ok writer.c:371: Test failed: 2272: got 11 expected 10 writer.c:373: Test failed: 2272: got <t>:E-3</t> expected <t>1E-2</t> writer.c:371: Test failed: 2272: got 12 expected 11 writer.c:373: Test failed: 2272: got <t>-:E-3</t> expected <t>-1E-2</t> writer.c:373: Test failed: 2272: got <t>1.7976931348623156E+308</t> expected <t>1.7976931348623157E+308</t> writer.c:373: Test failed: 2272: got <t>-1.7976931348623156E+308</t> expected <t>-1.7976931348623157E+308</t> writer.c:2316: Test failed: got 00000000 writer.c:2332: Test failed: got 00000000 Filed bug 49197 for that. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 Austin English <austinenglish(a)gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Fixed by SHA1| |6d70b1a66e01315dedf27d65738 | |b8f050552d7ed Resolution|--- |FIXED --- Comment #11 from Austin English <austinenglish(a)gmail.com> --- Fixed by 6d70b1a66e01315dedf27d65738b8f050552d7ed -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=43959 Alexandre Julliard <julliard(a)winehq.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED --- Comment #12 from Alexandre Julliard <julliard(a)winehq.org> --- Closing bugs fixed in 5.9. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
participants (2)
-
wine-bugs＠winehq.org -
WineHQ Bugzilla