> On 11-10-2016, at 5:47 PM, David T. Lewis <[hidden email]> wrote: > [snip] > I'll try to double-check it this weekend, but we already know that the > the version that I want people to use (i.e. mine) works fine. Granted > we did not try regenerating the C source from an oscog image, but I can't > think of any reason why that would not work. If we’re talking moving your slang code across then the only likely problems would be some places where slang meta-commands clash - IIRC #ifCpp:type stuff differs a bit. > > I really would prefer if people would use the original version of OSPP > that I maintain. After all, I tend to be the person who ends up answering > questions like "why doesn't OSProcess work on my Raspberry Pi?” I tend to agree. What do we need to do to make it so? > > Mind you, I'm not at all sorry that I bought that Raspberry Pi to figure > out why OSProcess was not working. Very cool little gadget, and I'm glad > you finally gave me an excuse to go get one :-) Much fun can be obtained for little money. Why, you can now get Flash under Chromium and run Scratch2 directly from MIT’s website. And … it’s…. sl….ow…. with odd dd… jum…ps. Seriously. Flash. <head-desk> > > Remind me to find a picture of the crude ultralight airplane I used > to fly back in the 1980's. It would make your brother look reasonably > sane. Inexplicably, I survived that one too. Just so long as the mudguard stays stayed in pace, eh? I’ve just built a latest vm with the unixosprocess .c file replaced by your one (taken from the old squeakvm.org tree, dated 6th jun, 8:10pm, hope that is the latest?) and loaded the OSProcess head package via squeakmap. I get 80 run, 71 passes, 0 expected failures, 5 failures, 4 errors, 0 unexpected passes. All the problems appear to be related to locking stuff, which is about as far as my ability will take me. Err, mind you 3 failures and 2 errors look like win32 tests, so just maybe they don’t count. tim -- tim Rowledge; [hidden email]; http://www.rowledge.org/tim Fractured Idiom:- POSH MORTEM - Death styles of the rich and famous |
On Wed, Oct 12, 2016 at 03:12:40PM -0700, tim Rowledge wrote: > > I???ve just built a latest vm with the unixosprocess .c file replaced by your one (taken from the old squeakvm.org tree, dated 6th jun, 8:10pm, hope that is the latest?) and loaded the OSProcess head package via squeakmap. I get 80 run, 71 passes, 0 expected failures, 5 failures, 4 errors, 0 unexpected passes. All the problems appear to be related to locking stuff, which is about as far as my ability will take me. Err, mind you 3 failures and 2 errors look like win32 tests, so just maybe they don???t count. > That will have to be "good enough" for now, sorry about that. I also get failures when running on Ubuntu on a fast machine. The tests have timing issues related to cooperating Squeak images that run in the tests, and the formerly all-green tests are failing now on the faster machines. I also suspect some differences related to file I/O differences as the OS runtimes have changed, but I can't say that for sure. In any case, your plugin is working, and OSProcess and CommandShell should work fine in normal day-to-day use. Dave |
In reply to this post by David T. Lewis
Hi David, Hi Tim,
On Mon, Oct 10, 2016 at 4:52 PM, David T. Lewis <[hidden email]> wrote:
This is unsatisfactory. I forked from VMConstruction-Plugins- In no way was I trying to undermine OSProcessPlugin or your authorship. i merely wanted to keep the code moving along with the rest of VMMaker.oscog so I could continue to generate the plugin source. So, in the absence of our harmonising the source so there is again only one VMConstruction-Plugins-
_,,,^..^,,,_ best, Eliot |
On Tue, Oct 18, 2016 at 03:46:13PM -0700, Eliot Miranda wrote: > > Hi David, Hi Tim, > > On Mon, Oct 10, 2016 at 4:52 PM, David T. Lewis <[hidden email]> wrote: > > > > > On Mon, Oct 10, 2016 at 11:03:16AM -0700, tim Rowledge wrote: > > > > > > > > > > On 10-10-2016, at 5:15 AM, Henrik Johansen < > > [hidden email]> wrote: > > > > > > > > 20160915 had one redeeming feature from my PoV; it shipped with a > > working OSProcessPlugin (thanks!) > > > > > > OK, that???s an interesting issue. I don???t recall adding it and I > > don???t recall removing it.. > > > > > > > Use VMConstruction-Plugins-OSProcessPlugin-dtl.40. That is the latest > > version of my OSProcessPlugin, and it works correctly on Raspbian. > > > > VMConstruction-Plugins-OSProcessPlugin.oscog-eem.53 does not work on > > Raspbian, and it fails UnixProcessAccessorTestCase>>testRedirectStdOutTo. > > > > This is unsatisfactory. I forked from > VMConstruction-Plugins-OSProcessPlugin-dtl.40 > because the Slang support has evolved for VMMaker.oscog, and not > sufficiently for VMMaker. For example, the VMMaker code assumes the source > is generated for the current platform, whereas the VMMaker.oscog code > assumes the plugin is generated irrespective of platform (because in > opensmalltalk/vm we store the generated code, and hence we want code > generated for all platforms, not just the "current" one). > > In no way was I trying to undermine OSProcessPlugin or your authorship. i > merely wanted to keep the code moving along with the rest of VMMaker.oscog > so I could continue to generate the plugin source. > > So, in the absence of our harmonising the source so there is again only one > VMConstruction-Plugins-OSProcessPlugin, can we please fold back the bug > fixes? If you point out what the bug fixes are I'll happily do the merge. > Hi Eliot, Sorry, but I do not know why the oscog version is crashing on Raspbian. Dave |
Free forum by Nabble | Edit this page |