From: Oren Ben-K. <or...@ri...> - 2002-09-04 08:09:40
|
Brian Ingerson [mailto:in...@tt...] wrote: > We did the #TAB thing because we were fairly sure people > would want it. But what if they don't? Then we'd forever have > to support something that we really didn't need to. Well... The problem is that we can't tell in advance. Sure, all users of YAML _now_ are "early adopters", have smart editors, configure their editors and so on. They definitely need no #TAB directive. The question is what happens when YAML starts to be widely adopted - that is, used by people that don't match the above profile. I don't know how much of a problem tabs would be then; nobody knows. The current spec "plays it safe" at the cost of implementation complexity. As you point out, we risk this complexity being paid unnecessarily. The converse risk is that we have a problem with "old" YAML implementations that don't know about #TAB. Which risk should we choose? > I'm in favor of taking the strong stance up front, and then > using #TAB down the road if we have to. Hmmm. I guess "Last Call" is a good time to raise all such nagging issues. I think we should get Clark on-line before we can make a decision, however - likewise with the timeperiod issue. Have fun, Oren Ben-Kiki |
From: Steve H. <sh...@zi...> - 2002-09-04 12:58:15
|
----- Original Message ----- From: "Oren Ben-Kiki" <or...@ri...> > > The current spec "plays it safe" at the cost of implementation complexity. > As you point out, we risk this complexity being paid unnecessarily. The > converse risk is that we have a problem with "old" YAML implementations that > don't know about #TAB. Which risk should we choose? > Let's risk going with the simplest approach. No tabs, no tab directive. -- Steve |
From: Rolf V. <rol...@he...> - 2002-09-04 15:24:49
|
Steve Howell wrote: > Let's risk going with the simplest approach. No tabs, no tab > directive. +1 Rolf |
From: Clark C . E. <cc...@cl...> - 2002-09-05 04:51:03
|
On Wed, Sep 04, 2002 at 08:58:42AM -0400, Steve Howell wrote: | From: "Oren Ben-Kiki" <or...@ri...> | > The current spec "plays it safe" at the cost of implementation complexity. | > As you point out, we risk this complexity being paid unnecessarily. The | > converse risk is that we have a problem with "old" YAML implementations that | > don't know about #TAB. Which risk should we choose? | | Let's risk going with the simplest approach. No tabs, no tab directive. I'm for this approach... we don't actually allow any-ole encoding. Clark |
From: Rolf V. <rol...@he...> - 2002-09-05 08:05:59
|
Steve Howell wrote: > Let's risk going with the simplest approach. No tabs, no tab directive. +1 Rolf |
From: Pete T. <pr...@vi...> - 2002-09-05 03:40:59
|
On Wed, 4 Sep 2002, Oren Ben-Kiki wrote: > Well... The problem is that we can't tell in advance. Sure, all > users of YAML _now_ are "early adopters", have smart editors, > configure their editors and so on. They definitely need no #TAB > directive. The question is what happens when YAML starts to be > widely adopted - that is, used by people that don't match the above > profile. I don't know how much of a problem tabs would be then; > nobody knows. One newbie's perspective: I've been using PyYaml for a few weeks now. I'm an emacs user from way back, but I jump from Windows to CygWin to various flavors of Linux throughout the day. I have no idea how my tabs are set on these various machines. I have learned that when PyYaml gives me a baffling error message, the first step is to untabify the file and try again. Out of a half dozen such episodes I can't remember a time when untabify didn't fix the problem. Other than a genuine bug (a # in a literal gets dismissed as a comment--I alerted Steve already), my tab stumbles have been the only negative part of using Yaml. I would much rather have a "no tabs please" message than to have the variety of other messages that can result from indentation being off. -- Pete |
From: Steve H. <sh...@zi...> - 2002-09-05 16:26:03
|
----- Original Message ----- From: "Pete TerMaat" <pr...@vi...> > On Wed, 4 Sep 2002, Oren Ben-Kiki wrote: > > Well... The problem is that we can't tell in advance. Sure, all > > users of YAML _now_ are "early adopters", have smart editors, > > configure their editors and so on. They definitely need no #TAB > > directive. The question is what happens when YAML starts to be > > widely adopted - that is, used by people that don't match the above > > profile. I don't know how much of a problem tabs would be then; > > nobody knows. > > One newbie's perspective: I've been using PyYaml for a few weeks now. > I'm an emacs user from way back, but I jump from Windows to CygWin to > various flavors of Linux throughout the day. I have no idea how my > tabs are set on these various machines. I have learned that when > PyYaml gives me a baffling error message, the first step is to > untabify the file and try again. Out of a half dozen such episodes I > can't remember a time when untabify didn't fix the problem. Other > than a genuine bug (a # in a literal gets dismissed as a comment--I > alerted Steve already), my tab stumbles have been the only negative > part of using Yaml. I would much rather have a "no tabs please" > message than to have the variety of other messages that can result > from indentation being off. > Yes, I will add it to my list to write a stern, clear message when I detect tabs. Thanks for the reminder on the comment bug, too. |
From: Brian I. <in...@tt...> - 2002-09-05 18:31:35
|
On 05/09/02 12:26 -0400, Steve Howell wrote: > ----- Original Message ----- > From: "Pete TerMaat" <pr...@vi...> > > On Wed, 4 Sep 2002, Oren Ben-Kiki wrote: > > > Well... The problem is that we can't tell in advance. Sure, all > > > users of YAML _now_ are "early adopters", have smart editors, > > > configure their editors and so on. They definitely need no #TAB > > > directive. The question is what happens when YAML starts to be > > > widely adopted - that is, used by people that don't match the above > > > profile. I don't know how much of a problem tabs would be then; > > > nobody knows. > > > > One newbie's perspective: I've been using PyYaml for a few weeks now. > > I'm an emacs user from way back, but I jump from Windows to CygWin to > > various flavors of Linux throughout the day. I have no idea how my > > tabs are set on these various machines. I have learned that when > > PyYaml gives me a baffling error message, the first step is to > > untabify the file and try again. Out of a half dozen such episodes I > > can't remember a time when untabify didn't fix the problem. Other > > than a genuine bug (a # in a literal gets dismissed as a comment--I > > alerted Steve already), my tab stumbles have been the only negative > > part of using Yaml. I would much rather have a "no tabs please" > > message than to have the variety of other messages that can result > > from indentation being off. > > > > Yes, I will add it to my list to write a stern, clear message when I detect > tabs. Perhaps we could come up with a common error message to use for our implementations. This probably needs to be refactored to be concise yet clear, and maybe should have a reference to more info. Here is one to get the discussion rolling. Refactor at will please: Tab characters were detected in your indentation. Tabs are not allowed in the YAML data format. Please change them to the appropriate number of spaces. See http://www.yaml.org/tab-policy for a full explanation. Cheers, Brian |
From: Steve H. <sh...@zi...> - 2002-09-06 04:36:54
|
----- Original Message ----- From: "Brian Ingerson" <in...@tt...> > > Perhaps we could come up with a common error message to use for our > implementations. This probably needs to be refactored to be concise yet > clear, and maybe should have a reference to more info. Here is one to get the > discussion rolling. Refactor at will please: > > Tab characters were detected in your indentation. Tabs are not > allowed in the YAML data format. Please change them to the > appropriate number of spaces. See http://www.yaml.org/tab-policy for > a full explanation. > I like this message. A good way for us to stay consistent with our messages is to put them in TestingSuite/error.yml. Cheers, Steve |
From: Brian I. <in...@tt...> - 2002-09-06 08:07:49
|
On 06/09/02 00:37 -0400, Steve Howell wrote: > ----- Original Message ----- > From: "Brian Ingerson" <in...@tt...> > > > > Perhaps we could come up with a common error message to use for our > > implementations. This probably needs to be refactored to be concise yet > > clear, and maybe should have a reference to more info. Here is one to get the > > discussion rolling. Refactor at will please: > > > > Tab characters were detected in your indentation. Tabs are not > > allowed in the YAML data format. Please change them to the > > appropriate number of spaces. See http://www.yaml.org/tab-policy for > > a full explanation. > > > > I like this message. > > A good way for us to stay consistent with our messages is to put them in > TestingSuite/error.yml. Good idea. I'm not sure I'll want to be consistent in all messages, but I think this one's important enough to get perfect. It will be nice to see what other people are doing error-wise though. Cheers, Brian |