Re: [RFC] COSS removal from 3.0
Amos Jeffries wrote: I'm expecting to roll 3.0.STABLE9 sometime over the next 5 days. One update still to be done is the removal of COSS. I had planned on just dead-coding (disabling) it. But with the configure recursion being dynamic thats not easily possible. I'm currently considering dropping an #error abortion into the top of all COSS code files to kill any builds trying to use it. Anyone have a better way? drop the code entirely from 3.0? Done. Noisy #error added. Basic code paths tested, but some may remain so if any 3.0 build dies with an error about COSS being unstable please let me know the compile messages. Thanks. Amos -- Please use Squid 2.7.STABLE4 or 3.0.STABLE8
Re: [RFC] COSS removal from 3.0
On Fri, 2008-09-05 at 01:46 +1200, Amos Jeffries wrote: I'm expecting to roll 3.0.STABLE9 sometime over the next 5 days. One update still to be done is the removal of COSS. I had planned on just dead-coding (disabling) it. But with the configure recursion being dynamic thats not easily possible. I'm currently considering dropping an #error abortion into the top of all COSS code files to kill any builds trying to use it. Anyone have a better way? drop the code entirely from 3.0? I would recommend against changing source files -- it will make porting Squid2 (or Squid v3.1) fixes harder. I would leave everything as is but add a ./configure warning if coss is being enabled. Would that work? #error is a one-liner and does not affect patching. #if 0 / dead-codng would be a two-liner with the same non-affect. I went for a configure fix first thing, there are at least two dynamic loops where we would have to add the sub-string test and warning message. Seems to me more work than its worth to do and test such a configure update. Why do you want to remove COSS in the first place? I know it has bugs, but so do many other features that are not in high demand. Moreover, I do expect COSS or its replacement to become popular in Squid3 again in the foreseeable future (which may prompt us to write back-patches for Squid 3.0). Well, Henrik stated it was not supposed to be in 3.0 in the first place. Adrian asserts, and a few user attempts confirm that its unusably broken in 3.0, with none of us planning fixes anytime. I thought I'd look into the options. We are looking at least another 3-4 stables before 3.1 is ready. But no sign of fixes there either. Amos
Re: [RFC] COSS removal from 3.0
On Fri, 2008-09-05 at 23:12 +1200, Amos Jeffries wrote: On Fri, 2008-09-05 at 01:46 +1200, Amos Jeffries wrote: I'm expecting to roll 3.0.STABLE9 sometime over the next 5 days. One update still to be done is the removal of COSS. I had planned on just dead-coding (disabling) it. But with the configure recursion being dynamic thats not easily possible. I'm currently considering dropping an #error abortion into the top of all COSS code files to kill any builds trying to use it. Anyone have a better way? drop the code entirely from 3.0? I would recommend against changing source files -- it will make porting Squid2 (or Squid v3.1) fixes harder. I would leave everything as is but add a ./configure warning if coss is being enabled. Would that work? #error is a one-liner and does not affect patching. #if 0 / dead-codng would be a two-liner with the same non-affect. I went for a configure fix first thing, there are at least two dynamic loops where we would have to add the sub-string test and warning message. Seems to me more work than its worth to do and test such a configure update. Understood. I have no strong objections to the disabling method you prefer. Thank you, Alex.
Re: [RFC] COSS removal from 3.0
2008/9/4 Amos Jeffries [EMAIL PROTECTED]: I'm expecting to roll 3.0.STABLE9 sometime over the next 5 days. One update still to be done is the removal of COSS. I had planned on just dead-coding (disabling) it. But with the configure recursion being dynamic thats not easily possible. I'm currently considering dropping an #error abortion into the top of all COSS code files to kill any builds trying to use it. Anyone have a better way? drop the code entirely from 3.0? I think thats perfectly fine for now. Adrian
Re: [RFC] COSS removal from 3.0
On Fri, 2008-09-05 at 01:46 +1200, Amos Jeffries wrote: I'm expecting to roll 3.0.STABLE9 sometime over the next 5 days. One update still to be done is the removal of COSS. I had planned on just dead-coding (disabling) it. But with the configure recursion being dynamic thats not easily possible. I'm currently considering dropping an #error abortion into the top of all COSS code files to kill any builds trying to use it. Anyone have a better way? drop the code entirely from 3.0? I would recommend against changing source files -- it will make porting Squid2 (or Squid v3.1) fixes harder. I would leave everything as is but add a ./configure warning if coss is being enabled. Would that work? Why do you want to remove COSS in the first place? I know it has bugs, but so do many other features that are not in high demand. Moreover, I do expect COSS or its replacement to become popular in Squid3 again in the foreseeable future (which may prompt us to write back-patches for Squid 3.0). Cheers, Alex.