Re: [racket-users] JFYI, c-level assertion failures during build
Thanks! I've pushed a repair. At Sun, 26 Feb 2017 00:11:50 -0500, "'John Clements' via Racket Users" wrote: > > > On Feb 25, 2017, at 5:36 AM, Matthew Flatt wrote: > > > > Did a fresh checkout solve the problem? > > > > If not, can you show more context to clarify which packages are being > > built when the crash happens. (That is, the crash is possibly not task > > in "1", which is shown here.) Even better, making with `CPUS=1` would > > clearly identify the package where the compiler is failing. > > Here’s the tail of the build process running with CPUS=1: > > raco setup: in /gui-lib/scheme/gui > raco setup: in /scheme-lib/scheme/unit/lang > raco setup: in /scheme-lib/scheme/unit > raco setup: in /syntax-color-lib/syntax-color > raco setup: in /data-lib/data > raco setup: in /data-lib/data/private > raco setup: in /gui-lib/mrlib > raco setup: in /gui-lib/mrlib/hierlist > raco setup: in /scheme-lib/scheme/signature/lang > raco setup: in /tex-table > raco setup: in /pict-lib/pict/private > raco setup: in /pict-lib/pict > raco setup: in /gui-lib/mrlib/private > raco setup: in /gui-lib/racket/gui/lang > raco setup: in /gui-lib/racket > raco setup: in /2d-test/tests > raco setup: making: /WebIDE/web-ide > raco setup: making: /WebIDE/web-ide/bountsy > raco setup: making: /WebIDE/web-ide/bountsy/conf > raco setup: making: /WebIDE/web-ide/bountsy/htdocs > raco setup: making: /WebIDE/web-ide/c-diff > raco setup: making: /WebIDE/web-ide/evaluators > raco setup: in /mongodb/db/mongodb/orm > Assertion failed: (!(used && (pre_body->count == 1) && > pre_body->vars[0]->optimize.known_val && > ((Scheme_Type)(scheme_once_used_type) > == (Scheme_Type)(intptr_t)(pre_body->vars[0]->optimize.known_val)) & > 0x1)?(Scheme_Type)scheme_integer_type:((Scheme_Object > *)(pre_body->vars[0]->optimize.known_val))->type))) && ((Scheme_Once_Used > *)pre_body->vars[0]->optimize.known_val)->moved)), function optimize_lets, > file ../../../racket/gc2/../src/optimize.c, line 7986. > make[2]: *** [plain-in-place] Abort trap: 6 > make[1]: *** [cpus-in-place] Error 2 > make: *** [in-place] Error 2 > make CPUS=1 62.08s user 6.78s system 77% cpu 1:28.62 total > > So it looks like … the mongodb package might be exposing this error? > > I’m actually running a fresh build right now, and I’ll see how it finishes. > > … > > Wow! The clean build worked fine, and then running raco pkg install mongodb > actually triggered the message. So it looks like this might be easily > reproducible, by installing mongodb. FWIW, I believe that the underlying > mongodb package is not installed on this machine, which could make it easier > to reproduce. > > I checked pkg-build.racket-lang.org, to see whether the problem was > observable > there, and it looks like pkg-build is using 6.8. I suppose this actually > makes > sense, since most package authors care about whether their packages build > against the stable version. Sounds like this could be a scenario where, say, > a > once-a-week CI process might be useful? (That assumes that this error is > cross-platform, of course.) > > … > > I dug just a bit further, and it looks like the problem is associated with > the > file mongodb/db/mongodb/orm/struct.rkt . Attempting to compile this file > triggers the error: > > /tmp/racket/racket/bin/raco make ./struct.rkt > Assertion failed: (!(used && (pre_body->count == 1) && > pre_body->vars[0]->optimize.known_val && > ((Scheme_Type)(scheme_once_used_type) > == (Scheme_Type)(intptr_t)(pre_body->vars[0]->optimize.known_val)) & > 0x1)?(Scheme_Type)scheme_integer_type:((Scheme_Object > *)(pre_body->vars[0]->optimize.known_val))->type))) && ((Scheme_Once_Used > *)pre_body->vars[0]->optimize.known_val)->moved)), function optimize_lets, > file ../../../racket/gc2/../src/optimize.c, line 7986. > zsh: abort /tmp/racket/racket/bin/raco make ./struct.rkt > > It occurred to me that just maybe this could be due to something unsafe in > mongodb, but I did a quick grep for ‘ffi’ and ‘unsafe’, and in fact I don’t > think the package has any weirdness in it at all. > > John > > > > > > > > > > > At Fri, 24 Feb 2017 19:52:43 -0500, "'John Clements' via Racket Users" > > wrote: > >> I’m currently in a state where calling `make` in my top-level racket > directory > >> results in output ending thusly: > >> > >> … > >> raco setup: 1 making: /cldr-bcp47/cldr/bcp47/data > >> Assertion failed: (!(used && (pre_body->count == 1) && > >> pre_body->vars[0]->optimize.known_val && > ((Scheme_Type)(scheme_once_used_type) > >> == (Scheme_Type)(intptr_t)(pre_body->vars[0]->optimize.known_val)) & > >> 0x1)?(Scheme_Type)scheme_integer_type:((Scheme_Object > >> *)(pre_body->vars[0]->optimize.known_val))->type))) && ((Scheme_Once_Used > >> *)pre_body->vars[0]->optimize.known_val)->moved)), function optimize_lets, > >> file ../../../racket/gc2/../src/optimize.c, line 7986. > >> make[1]: *** [plain-in-place] Abo
Re: [racket-users] JFYI, c-level assertion failures during build
Snapshot-based pkg builds are here: https://plt.eecs.northwestern.edu/pkg-build/ Robby On Sat, Feb 25, 2017 at 11:11 PM, 'John Clements' via Racket Users wrote: > >> On Feb 25, 2017, at 5:36 AM, Matthew Flatt wrote: >> >> Did a fresh checkout solve the problem? >> >> If not, can you show more context to clarify which packages are being >> built when the crash happens. (That is, the crash is possibly not task >> in "1", which is shown here.) Even better, making with `CPUS=1` would >> clearly identify the package where the compiler is failing. > > Here’s the tail of the build process running with CPUS=1: > > raco setup: in /gui-lib/scheme/gui > raco setup: in /scheme-lib/scheme/unit/lang > raco setup: in /scheme-lib/scheme/unit > raco setup: in /syntax-color-lib/syntax-color > raco setup: in /data-lib/data > raco setup: in /data-lib/data/private > raco setup: in /gui-lib/mrlib > raco setup: in /gui-lib/mrlib/hierlist > raco setup: in /scheme-lib/scheme/signature/lang > raco setup: in /tex-table > raco setup: in /pict-lib/pict/private > raco setup: in /pict-lib/pict > raco setup: in /gui-lib/mrlib/private > raco setup: in /gui-lib/racket/gui/lang > raco setup: in /gui-lib/racket > raco setup: in /2d-test/tests > raco setup: making: /WebIDE/web-ide > raco setup: making: /WebIDE/web-ide/bountsy > raco setup: making: /WebIDE/web-ide/bountsy/conf > raco setup: making: /WebIDE/web-ide/bountsy/htdocs > raco setup: making: /WebIDE/web-ide/c-diff > raco setup: making: /WebIDE/web-ide/evaluators > raco setup: in /mongodb/db/mongodb/orm > Assertion failed: (!(used && (pre_body->count == 1) && > pre_body->vars[0]->optimize.known_val && > ((Scheme_Type)(scheme_once_used_type) == > (Scheme_Type)(intptr_t)(pre_body->vars[0]->optimize.known_val)) & > 0x1)?(Scheme_Type)scheme_integer_type:((Scheme_Object > *)(pre_body->vars[0]->optimize.known_val))->type))) && ((Scheme_Once_Used > *)pre_body->vars[0]->optimize.known_val)->moved)), function optimize_lets, > file ../../../racket/gc2/../src/optimize.c, line 7986. > make[2]: *** [plain-in-place] Abort trap: 6 > make[1]: *** [cpus-in-place] Error 2 > make: *** [in-place] Error 2 > make CPUS=1 62.08s user 6.78s system 77% cpu 1:28.62 total > > So it looks like … the mongodb package might be exposing this error? > > I’m actually running a fresh build right now, and I’ll see how it finishes. > > … > > Wow! The clean build worked fine, and then running raco pkg install mongodb > actually triggered the message. So it looks like this might be easily > reproducible, by installing mongodb. FWIW, I believe that the underlying > mongodb package is not installed on this machine, which could make it easier > to reproduce. > > I checked pkg-build.racket-lang.org, to see whether the problem was > observable there, and it looks like pkg-build is using 6.8. I suppose this > actually makes sense, since most package authors care about whether their > packages build against the stable version. Sounds like this could be a > scenario where, say, a once-a-week CI process might be useful? (That assumes > that this error is cross-platform, of course.) > > … > > I dug just a bit further, and it looks like the problem is associated with > the file mongodb/db/mongodb/orm/struct.rkt . Attempting to compile this file > triggers the error: > > /tmp/racket/racket/bin/raco make ./struct.rkt > Assertion failed: (!(used && (pre_body->count == 1) && > pre_body->vars[0]->optimize.known_val && > ((Scheme_Type)(scheme_once_used_type) == > (Scheme_Type)(intptr_t)(pre_body->vars[0]->optimize.known_val)) & > 0x1)?(Scheme_Type)scheme_integer_type:((Scheme_Object > *)(pre_body->vars[0]->optimize.known_val))->type))) && ((Scheme_Once_Used > *)pre_body->vars[0]->optimize.known_val)->moved)), function optimize_lets, > file ../../../racket/gc2/../src/optimize.c, line 7986. > zsh: abort /tmp/racket/racket/bin/raco make ./struct.rkt > > It occurred to me that just maybe this could be due to something unsafe in > mongodb, but I did a quick grep for ‘ffi’ and ‘unsafe’, and in fact I don’t > think the package has any weirdness in it at all. > > John > > > > > > > >> >> At Fri, 24 Feb 2017 19:52:43 -0500, "'John Clements' via Racket Users" wrote: >>> I’m currently in a state where calling `make` in my top-level racket >>> directory >>> results in output ending thusly: >>> >>> … >>> raco setup: 1 making: /cldr-bcp47/cldr/bcp47/data >>> Assertion failed: (!(used && (pre_body->count == 1) && >>> pre_body->vars[0]->optimize.known_val && >>> ((Scheme_Type)(scheme_once_used_type) >>> == (Scheme_Type)(intptr_t)(pre_body->vars[0]->optimize.known_val)) & >>> 0x1)?(Scheme_Type)scheme_integer_type:((Scheme_Object >>> *)(pre_body->vars[0]->optimize.known_val))->type))) && ((Scheme_Once_Used >>> *)pre_body->vars[0]->optimize.known_val)->moved)), function optimize_lets, >>> file ../../../racket/gc2/../src/optimize.c, line 7986. >>> make[1]: *** [plain-in-place] Abort trap: 6 >
Re: [racket-users] JFYI, c-level assertion failures during build
> On Feb 25, 2017, at 5:36 AM, Matthew Flatt wrote: > > Did a fresh checkout solve the problem? > > If not, can you show more context to clarify which packages are being > built when the crash happens. (That is, the crash is possibly not task > in "1", which is shown here.) Even better, making with `CPUS=1` would > clearly identify the package where the compiler is failing. Here’s the tail of the build process running with CPUS=1: raco setup: in /gui-lib/scheme/gui raco setup: in /scheme-lib/scheme/unit/lang raco setup: in /scheme-lib/scheme/unit raco setup: in /syntax-color-lib/syntax-color raco setup: in /data-lib/data raco setup: in /data-lib/data/private raco setup: in /gui-lib/mrlib raco setup: in /gui-lib/mrlib/hierlist raco setup: in /scheme-lib/scheme/signature/lang raco setup: in /tex-table raco setup: in /pict-lib/pict/private raco setup: in /pict-lib/pict raco setup: in /gui-lib/mrlib/private raco setup: in /gui-lib/racket/gui/lang raco setup: in /gui-lib/racket raco setup: in /2d-test/tests raco setup: making: /WebIDE/web-ide raco setup: making: /WebIDE/web-ide/bountsy raco setup: making: /WebIDE/web-ide/bountsy/conf raco setup: making: /WebIDE/web-ide/bountsy/htdocs raco setup: making: /WebIDE/web-ide/c-diff raco setup: making: /WebIDE/web-ide/evaluators raco setup: in /mongodb/db/mongodb/orm Assertion failed: (!(used && (pre_body->count == 1) && pre_body->vars[0]->optimize.known_val && ((Scheme_Type)(scheme_once_used_type) == (Scheme_Type)(intptr_t)(pre_body->vars[0]->optimize.known_val)) & 0x1)?(Scheme_Type)scheme_integer_type:((Scheme_Object *)(pre_body->vars[0]->optimize.known_val))->type))) && ((Scheme_Once_Used *)pre_body->vars[0]->optimize.known_val)->moved)), function optimize_lets, file ../../../racket/gc2/../src/optimize.c, line 7986. make[2]: *** [plain-in-place] Abort trap: 6 make[1]: *** [cpus-in-place] Error 2 make: *** [in-place] Error 2 make CPUS=1 62.08s user 6.78s system 77% cpu 1:28.62 total So it looks like … the mongodb package might be exposing this error? I’m actually running a fresh build right now, and I’ll see how it finishes. … Wow! The clean build worked fine, and then running raco pkg install mongodb actually triggered the message. So it looks like this might be easily reproducible, by installing mongodb. FWIW, I believe that the underlying mongodb package is not installed on this machine, which could make it easier to reproduce. I checked pkg-build.racket-lang.org, to see whether the problem was observable there, and it looks like pkg-build is using 6.8. I suppose this actually makes sense, since most package authors care about whether their packages build against the stable version. Sounds like this could be a scenario where, say, a once-a-week CI process might be useful? (That assumes that this error is cross-platform, of course.) … I dug just a bit further, and it looks like the problem is associated with the file mongodb/db/mongodb/orm/struct.rkt . Attempting to compile this file triggers the error: /tmp/racket/racket/bin/raco make ./struct.rkt Assertion failed: (!(used && (pre_body->count == 1) && pre_body->vars[0]->optimize.known_val && ((Scheme_Type)(scheme_once_used_type) == (Scheme_Type)(intptr_t)(pre_body->vars[0]->optimize.known_val)) & 0x1)?(Scheme_Type)scheme_integer_type:((Scheme_Object *)(pre_body->vars[0]->optimize.known_val))->type))) && ((Scheme_Once_Used *)pre_body->vars[0]->optimize.known_val)->moved)), function optimize_lets, file ../../../racket/gc2/../src/optimize.c, line 7986. zsh: abort /tmp/racket/racket/bin/raco make ./struct.rkt It occurred to me that just maybe this could be due to something unsafe in mongodb, but I did a quick grep for ‘ffi’ and ‘unsafe’, and in fact I don’t think the package has any weirdness in it at all. John > > At Fri, 24 Feb 2017 19:52:43 -0500, "'John Clements' via Racket Users" wrote: >> I’m currently in a state where calling `make` in my top-level racket >> directory >> results in output ending thusly: >> >> … >> raco setup: 1 making: /cldr-bcp47/cldr/bcp47/data >> Assertion failed: (!(used && (pre_body->count == 1) && >> pre_body->vars[0]->optimize.known_val && >> ((Scheme_Type)(scheme_once_used_type) >> == (Scheme_Type)(intptr_t)(pre_body->vars[0]->optimize.known_val)) & >> 0x1)?(Scheme_Type)scheme_integer_type:((Scheme_Object >> *)(pre_body->vars[0]->optimize.known_val))->type))) && ((Scheme_Once_Used >> *)pre_body->vars[0]->optimize.known_val)->moved)), function optimize_lets, >> file ../../../racket/gc2/../src/optimize.c, line 7986. >> make[1]: *** [plain-in-place] Abort trap: 6 >> make: *** [in-place] Error 2 >> >> >> My next step is to try a fresh checkout… any idea what’s going on here? >> >> John Clements >> >> >> >> -- >> You received this message because you are subscribed to the Google Groups >> "Racket Users" group. >> To unsubscribe from this group and stop receiving emails from it, send an >
Re: [racket-users] JFYI, c-level assertion failures during build
Did a fresh checkout solve the problem? If not, can you show more context to clarify which packages are being built when the crash happens. (That is, the crash is possibly not task in "1", which is shown here.) Even better, making with `CPUS=1` would clearly identify the package where the compiler is failing. At Fri, 24 Feb 2017 19:52:43 -0500, "'John Clements' via Racket Users" wrote: > I’m currently in a state where calling `make` in my top-level racket > directory > results in output ending thusly: > > … > raco setup: 1 making: /cldr-bcp47/cldr/bcp47/data > Assertion failed: (!(used && (pre_body->count == 1) && > pre_body->vars[0]->optimize.known_val && > ((Scheme_Type)(scheme_once_used_type) > == (Scheme_Type)(intptr_t)(pre_body->vars[0]->optimize.known_val)) & > 0x1)?(Scheme_Type)scheme_integer_type:((Scheme_Object > *)(pre_body->vars[0]->optimize.known_val))->type))) && ((Scheme_Once_Used > *)pre_body->vars[0]->optimize.known_val)->moved)), function optimize_lets, > file ../../../racket/gc2/../src/optimize.c, line 7986. > make[1]: *** [plain-in-place] Abort trap: 6 > make: *** [in-place] Error 2 > > > My next step is to try a fresh checkout… any idea what’s going on here? > > John Clements > > > > -- > You received this message because you are subscribed to the Google Groups > "Racket Users" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to racket-users+unsubscr...@googlegroups.com. > For more options, visit https://groups.google.com/d/optout. -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to racket-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.