A couple of my SqueakMap entries (maybe more, I have not checked them all)
have decided to stop working. I got the AioPlugin entry working again by re-saving the entry, but XDisplayControlPlugin is screwed up and I can't seem to convince it to start working again. What the user sees from a SqueakMap Package Loader is a debugger with the rather unhelpful "Error: can't find EOCD position" message. <rant> How many other SqueakMap entries are messed up and the authors don't even know it? And what is a person supposed to think when they see this? It makes people think that AioPlugin and XDisplayControlPlugin are broken, unmaintained, and unavailable, none of which is true. </rant> Dave |
David T. Lewis skrev:
> A couple of my SqueakMap entries (maybe more, I have not checked them all) > have decided to stop working. I got the AioPlugin entry working again by > re-saving the entry, but XDisplayControlPlugin is screwed up and I can't > seem to convince it to start working again. > > What the user sees from a SqueakMap Package Loader is a debugger with > the rather unhelpful "Error: can't find EOCD position" message. > > <rant> > How many other SqueakMap entries are messed up and the authors don't > even know it? And what is a person supposed to think when they see this? > It makes people think that AioPlugin and XDisplayControlPlugin are broken, > unmaintained, and unavailable, none of which is true. > </rant> > > Dave > > > > I get things to work on windows with it. It also fixes an issue when trying to access www.linuxfr.org from Scamper. I have not read trough the method to see what have changed that cause the bug. Karl HTTPSocket class-httpGetDocumentargsacceptrequest.st.gz (2K) Download Attachment |
On Sun, Oct 08, 2006 at 05:40:27PM +0200, karl wrote:
> David T. Lewis skrev: > >A couple of my SqueakMap entries (maybe more, I have not checked them all) > >have decided to stop working. I got the AioPlugin entry working again by > >re-saving the entry, but XDisplayControlPlugin is screwed up and I can't > >seem to convince it to start working again. > > > >What the user sees from a SqueakMap Package Loader is a debugger with > >the rather unhelpful "Error: can't find EOCD position" message. > > > ><rant> > >How many other SqueakMap entries are messed up and the authors don't > >even know it? And what is a person supposed to think when they see this? > >It makes people think that AioPlugin and XDisplayControlPlugin are broken, > >unmaintained, and unavailable, none of which is true. > ></rant> > > > I have found that there is a bug in HttpSocket. > Try the attached method which is a older version of the method that fails. > I get things to work on windows with it. It also fixes an issue when > trying to > access www.linuxfr.org from Scamper. I have not read trough the method > to see what have changed that cause the bug. > Karl Thanks Karl, I tried the patch you sent, but this seems to be a different problem. I think SqueakMap keeps some sort of validity checks on the file links and/or its internal caches, which apparently get screwed up at random times for no apparent reason. Oh, and I was wrong, I did not get the AioPlugin entry working, it's just broken differently now. I probably should not have tried to "fix" it. Dave |
David T. Lewis skrev:
> On Sun, Oct 08, 2006 at 05:40:27PM +0200, karl wrote: > >> David T. Lewis skrev: >> >>> A couple of my SqueakMap entries (maybe more, I have not checked them all) >>> have decided to stop working. I got the AioPlugin entry working again by >>> re-saving the entry, but XDisplayControlPlugin is screwed up and I can't >>> seem to convince it to start working again. >>> >>> What the user sees from a SqueakMap Package Loader is a debugger with >>> the rather unhelpful "Error: can't find EOCD position" message. >>> >>> <rant> >>> How many other SqueakMap entries are messed up and the authors don't >>> even know it? And what is a person supposed to think when they see this? >>> It makes people think that AioPlugin and XDisplayControlPlugin are broken, >>> unmaintained, and unavailable, none of which is true. >>> </rant> >>> >>> >> I have found that there is a bug in HttpSocket. >> Try the attached method which is a older version of the method that fails. >> I get things to work on windows with it. It also fixes an issue when >> trying to >> access www.linuxfr.org from Scamper. I have not read trough the method >> to see what have changed that cause the bug. >> Karl >> > > Thanks Karl, > > I tried the patch you sent, but this seems to be a different problem. > I think SqueakMap keeps some sort of validity checks on the file links > and/or its internal caches, which apparently get screwed up at random > times for no apparent reason. Oh, and I was wrong, I did not get the > AioPlugin entry working, it's just broken differently now. I probably > should not have tried to "fix" it. > > Dave > > > > HTTPSocket version and downloads a bogus file. Look at the issue brought up by Göran before.(attached) Karl [hidden email] skrev: > Hi! > > Benoit St-Jean <[hidden email]> wrote: > >> Oups! Posted my reply in French... >> >> Stéphane, I never implied that this release was s***. I noticed the >> huge number of changes that went into >> this release and good job to everyone. BUT right now >> a lot of packages don't load into 3.9. >> >> I guess those packages will eventually get updated but >> right now, this is very frustrating! >> >> For instance, on my last attempt trying to install the >> Chronos package creates a "Error occurred during >> install: can't find EOCD position"... >> > > That particular one sounds like SM server side cache gone wrong. Let me > try to fix that one right away.... > > ...churn, churn, blonk, klonk... silence... Hmm? twirl, bang! bang! > Sigh... > > Hmmm, ok, I think this was harder than I thought. It seems to me that > the stuff returned from that URL is this: > > '<HTML> > <HEAD> > <META HTTP-EQUIV="Content-Type" CONTENT="text/html; > charset=iso-8859-1"> > <TITLE>chronos-st.org:80</TITLE> > </HEAD> > <FRAMESET > ROWS="130,*" FRAMEBORDER=0 BORDER=0 FRAMESPACING=0> > <FRAME > NAME="content" > SRC="http://vhost.r4l.com/top_frm?domain=chronos-st.org:80" > marginheight=0 marginwidth=0 noresize scrolling=no> > <FRAME > NAME="content" > SRC="http://apps5.oingo.com/apps/domainpark/domainpark.cgi?s=chronos-st. > org:80&cid=REGI991N&uid=62.181.231.43&dp_p4pid=oingo_inclusion_xml_10&dp > _lp=7&dp_pu=2&dp_format=1.3" marginheight=0 marginwidth=0 noresize> > > <NOFRAMES> > <META http-equiv="Content-Type" content="text/html; > charset=iso-8859-1"> > <META http-equiv="refresh" content="0; > url=chronos-st.org:80"> > </NOFRAMES> > </FRAMESET> > </HTML> > ' > > But using say wget or firefox I get the Chronos.sar file allright. Not > sure why this code does not work: > > 'http://chronos-st.org/downloads/Squeak/Chronos.sar' asUrl > retrieveContents contents > > Anyone? > > regards, Göran > > > HTTPSocket class httpGetDocument: url args: args accept: mimeType request: requestString to version tlk 1/22/2004 and I got the right content downloaded. Karl |
In reply to this post by karl-8
Please file a bug in Mantis on this.
On Oct 8, 2006, at 8:40 AM, karl wrote: > I have found that there is a bug in HttpSocket. > Try the attached method which is a older version of the method that > fails. > I get things to work on windows with it. It also fixes an issue > when trying to > access www.linuxfr.org from Scamper. I have not read trough the method > to see what have changed that cause the bug. > Karl > <HTTPSocket class-httpGetDocumentargsacceptrequest.st.gz> > |
Free forum by Nabble | Edit this page |