Re: GSoC: Advanced Find and Replace Questions

[Albus X]

Thanks for your feedback!

I have updated my proposal again, trying to explain everything clearer.

I thought that implement regular expression parsing from scratch is
not difficult, and can lead to a more uniform structure. But after
some consideration, I find it not worth maintaining a database of
character classes inside Lyx, Therefore I modified my proposal, and
the whole procedure is now simpler.

On Wed, May 1, 2013 at 4:14 AM, Tommaso Cucinotta tomm...@lyx.org wrote:
 On 30/04/13 03:11, Albus X wrote:
 My proposal is updated. I removed the mentioning of Prolog, made the
 algorithm description clearer, and added more explanation.

 Hi,

 I could spot a few vague/ambiguous parts in the current description:

 - what do you mean by variable name ?
 - what does it mean to drop in the result all unnamed variables ?
 - what's wrong with using parentheses within regexps, and referring
   them back within the search pattern itself, as well as within the
   text to be replaced ?
 - how would you plan to realize, for example, the option you describe
   to treat all parentheses as the same ?
 - what regexp features would you plan to support ? can you provide
   a few examples / regexps that you think should work, and how ?
 - in section Extending..., .1: Preproces... into a list, that seems
   to clash with the description above about the matching algorithm;
   perhaps you may want to clarify ?

 Thanks,

 T.

Re: GSoC: Advanced Find and Replace Questions

[Albus X]

Thanks for your feedback!

I have updated my proposal again, trying to explain everything clearer.

I thought that implement regular expression parsing from scratch is
not difficult, and can lead to a more uniform structure. But after
some consideration, I find it not worth maintaining a database of
character classes inside Lyx, Therefore I modified my proposal, and
the whole procedure is now simpler.

On Wed, May 1, 2013 at 4:14 AM, Tommaso Cucinotta  wrote:
> On 30/04/13 03:11, Albus X wrote:
>> My proposal is updated. I removed the mentioning of Prolog, made the
>> algorithm description clearer, and added more explanation.
>
> Hi,
>
> I could spot a few vague/ambiguous parts in the current description:
>
> - what do you mean by "variable name" ?
> - what does it mean to "drop in the result all unnamed variables" ?
> - what's wrong with using parentheses within regexps, and referring
>   them back within the search pattern itself, as well as within the
>   text to be replaced ?
> - how would you plan to realize, for example, the option you describe
>   to treat all parentheses as the same ?
> - what regexp features would you plan to support ? can you provide
>   a few examples / regexps that you think should work, and how ?
> - in section "Extending...", .1: Preproces... into a list, that seems
>   to clash with the description above about the matching algorithm;
>   perhaps you may want to clarify ?
>
> Thanks,
>
> T.
>

Re: GSoC: Advanced Find and Replace Questions

[Tommaso Cucinotta]

On 30/04/13 03:11, Albus X wrote:
 My proposal is updated. I removed the mentioning of Prolog, made the
 algorithm description clearer, and added more explanation.

Hi,

I could spot a few vague/ambiguous parts in the current description:

- what do you mean by variable name ?
- what does it mean to drop in the result all unnamed variables ?
- what's wrong with using parentheses within regexps, and referring
  them back within the search pattern itself, as well as within the
  text to be replaced ?
- how would you plan to realize, for example, the option you describe
  to treat all parentheses as the same ?
- what regexp features would you plan to support ? can you provide
  a few examples / regexps that you think should work, and how ?
- in section Extending..., .1: Preproces... into a list, that seems
  to clash with the description above about the matching algorithm;
  perhaps you may want to clarify ?

Thanks,

T.

Re: GSoC: Advanced Find and Replace Questions

[Tommaso Cucinotta]

On 30/04/13 03:11, Albus X wrote:
> My proposal is updated. I removed the mentioning of Prolog, made the
> algorithm description clearer, and added more explanation.

Hi,

I could spot a few vague/ambiguous parts in the current description:

- what do you mean by "variable name" ?
- what does it mean to "drop in the result all unnamed variables" ?
- what's wrong with using parentheses within regexps, and referring
  them back within the search pattern itself, as well as within the
  text to be replaced ?
- how would you plan to realize, for example, the option you describe
  to treat all parentheses as the same ?
- what regexp features would you plan to support ? can you provide
  a few examples / regexps that you think should work, and how ?
- in section "Extending...", .1: Preproces... into a list, that seems
  to clash with the description above about the matching algorithm;
  perhaps you may want to clarify ?

Thanks,

T.

Re: GSoC: Advanced Find and Replace Questions

[Albus X]

I have officially submitted my proposal at
https://google-melange.appspot.com/gsoc/proposal/review/google/gsoc2013/albuseer/1

Please have a review.

On Mon, Apr 29, 2013 at 10:18 AM, Albus X albus...@gmail.com wrote:
 I am not very familiar with git, what is the meaning of
 [712e1871/lyxgit]? I just cloned the source now.
 I got it. It's the commit ID and I finally find out what your changed.

 On Mon, Apr 29, 2013 at 8:49 AM, Albus X albus...@gmail.com wrote:
 On Mon, Apr 29, 2013 at 7:15 AM, Tommaso Cucinotta tomm...@lyx.org wrote:
 Interestingly, those few scenarios which are documented in the manual, are
 not really included in the automated test cases as from 
 devel/autotests/findadv-*.in.

 In fact, it's not in the manual. The manual said I can search by
 regex, and gave that example, but not said about how to use the regex
 box. Maybe should add some explanation to the manual?

 these are all nice observations, but they have to be raised on the list with
 some different subject line, e.g., Usability issues in math-writing or 
 similar,
 otherwise you won't get the necessary attention by the right people.

 Thanks for your pointing out.

 At least, now there's a regression test for this [712e1871/lyxgit].

 I am not very familiar with git, what is the meaning of
 [712e1871/lyxgit]? I just cloned the source now.

 Thanks.

Re: GSoC: Advanced Find and Replace Questions

[Cyrille Artho]

Hi Albus,
Thank you for your proposal.
I am not very familiar with the project, but I see that you mention Prolog 
in your proposal. Are you planning on using Prolog? Do you have any ideas 
in how a similar algorithm could be implemented in C or C++ instead? (It 
would be good if a LyX user did not have to install Prolog as LyX currently 
does not depend on it.)


Albus X wrote:

I have officially submitted my proposal at
https://google-melange.appspot.com/gsoc/proposal/review/google/gsoc2013/albuseer/1

Please have a review.

On Mon, Apr 29, 2013 at 10:18 AM, Albus X albus...@gmail.com wrote:

I am not very familiar with git, what is the meaning of
[712e1871/lyxgit]? I just cloned the source now.

I got it. It's the commit ID and I finally find out what your changed.

On Mon, Apr 29, 2013 at 8:49 AM, Albus X albus...@gmail.com wrote:

On Mon, Apr 29, 2013 at 7:15 AM, Tommaso Cucinotta tomm...@lyx.org wrote:

Interestingly, those few scenarios which are documented in the manual, are
not really included in the automated test cases as from 
devel/autotests/findadv-*.in.


In fact, it's not in the manual. The manual said I can search by
regex, and gave that example, but not said about how to use the regex
box. Maybe should add some explanation to the manual?


these are all nice observations, but they have to be raised on the list with
some different subject line, e.g., Usability issues in math-writing or 
similar,
otherwise you won't get the necessary attention by the right people.


Thanks for your pointing out.


At least, now there's a regression test for this [712e1871/lyxgit].


I am not very familiar with git, what is the meaning of
[712e1871/lyxgit]? I just cloned the source now.

Thanks.


--
Regards,
Cyrille Artho - http://artho.com/
Until Eve arrived, this was a man's world.
-- Richard Armour

Re: GSoC: Advanced Find and Replace Questions

[Albus X]

No. I am indeed not familiar with programming in Prolog. I mentioned
it because the pattern matching feature of Prolog, (or maybe many
functional languages?) is implemented that way. I think find and
replace is also a kind of pattern matching, and for the replace
feature, the substitution table calculated in that algorithm does the
exact job for group referencing in regexp.

In the last one year and more, I focused mainly on the topics of
compilers and programming language design. Therefore this project
immediately remind me of that algorithm. I have indeed an description
of the algorithm in my proposal, therefore it can be implemented
easily in C++. Maybe my description is not clear enough, I will add
more detail to it now.

On Tue, Apr 30, 2013 at 8:44 AM, Cyrille Artho c.ar...@aist.go.jp wrote:
 Hi Albus,
 Thank you for your proposal.
 I am not very familiar with the project, but I see that you mention Prolog
 in your proposal. Are you planning on using Prolog? Do you have any ideas in
 how a similar algorithm could be implemented in C or C++ instead? (It would
 be good if a LyX user did not have to install Prolog as LyX currently does
 not depend on it.)


 Albus X wrote:

 I have officially submitted my proposal at

 https://google-melange.appspot.com/gsoc/proposal/review/google/gsoc2013/albuseer/1

 Please have a review.

 On Mon, Apr 29, 2013 at 10:18 AM, Albus X albus...@gmail.com wrote:

 I am not very familiar with git, what is the meaning of
 [712e1871/lyxgit]? I just cloned the source now.

 I got it. It's the commit ID and I finally find out what your changed.

 On Mon, Apr 29, 2013 at 8:49 AM, Albus X albus...@gmail.com wrote:

 On Mon, Apr 29, 2013 at 7:15 AM, Tommaso Cucinotta tomm...@lyx.org
 wrote:

 Interestingly, those few scenarios which are documented in the manual,
 are
 not really included in the automated test cases as from
 devel/autotests/findadv-*.in.


 In fact, it's not in the manual. The manual said I can search by
 regex, and gave that example, but not said about how to use the regex
 box. Maybe should add some explanation to the manual?

 these are all nice observations, but they have to be raised on the list
 with
 some different subject line, e.g., Usability issues in math-writing
 or similar,
 otherwise you won't get the necessary attention by the right people.


 Thanks for your pointing out.

 At least, now there's a regression test for this [712e1871/lyxgit].


 I am not very familiar with git, what is the meaning of
 [712e1871/lyxgit]? I just cloned the source now.

 Thanks.


 --
 Regards,
 Cyrille Artho - http://artho.com/
 Until Eve arrived, this was a man's world.
 -- Richard Armour

Re: GSoC: Advanced Find and Replace Questions

[Albus X]

My proposal is updated. I removed the mentioning of Prolog, made the
algorithm description clearer, and added more explanation.

On Tue, Apr 30, 2013 at 8:58 AM, Albus X albus...@gmail.com wrote:
 No. I am indeed not familiar with programming in Prolog. I mentioned
 it because the pattern matching feature of Prolog, (or maybe many
 functional languages?) is implemented that way. I think find and
 replace is also a kind of pattern matching, and for the replace
 feature, the substitution table calculated in that algorithm does the
 exact job for group referencing in regexp.

 In the last one year and more, I focused mainly on the topics of
 compilers and programming language design. Therefore this project
 immediately remind me of that algorithm. I have indeed an description
 of the algorithm in my proposal, therefore it can be implemented
 easily in C++. Maybe my description is not clear enough, I will add
 more detail to it now.

 On Tue, Apr 30, 2013 at 8:44 AM, Cyrille Artho c.ar...@aist.go.jp wrote:
 Hi Albus,
 Thank you for your proposal.
 I am not very familiar with the project, but I see that you mention Prolog
 in your proposal. Are you planning on using Prolog? Do you have any ideas in
 how a similar algorithm could be implemented in C or C++ instead? (It would
 be good if a LyX user did not have to install Prolog as LyX currently does
 not depend on it.)


 Albus X wrote:

 I have officially submitted my proposal at

 https://google-melange.appspot.com/gsoc/proposal/review/google/gsoc2013/albuseer/1

 Please have a review.

 On Mon, Apr 29, 2013 at 10:18 AM, Albus X albus...@gmail.com wrote:

 I am not very familiar with git, what is the meaning of
 [712e1871/lyxgit]? I just cloned the source now.

 I got it. It's the commit ID and I finally find out what your changed.

 On Mon, Apr 29, 2013 at 8:49 AM, Albus X albus...@gmail.com wrote:

 On Mon, Apr 29, 2013 at 7:15 AM, Tommaso Cucinotta tomm...@lyx.org
 wrote:

 Interestingly, those few scenarios which are documented in the manual,
 are
 not really included in the automated test cases as from
 devel/autotests/findadv-*.in.


 In fact, it's not in the manual. The manual said I can search by
 regex, and gave that example, but not said about how to use the regex
 box. Maybe should add some explanation to the manual?

 these are all nice observations, but they have to be raised on the list
 with
 some different subject line, e.g., Usability issues in math-writing
 or similar,
 otherwise you won't get the necessary attention by the right people.


 Thanks for your pointing out.

 At least, now there's a regression test for this [712e1871/lyxgit].


 I am not very familiar with git, what is the meaning of
 [712e1871/lyxgit]? I just cloned the source now.

 Thanks.


 --
 Regards,
 Cyrille Artho - http://artho.com/
 Until Eve arrived, this was a man's world.
 -- Richard Armour

Re: GSoC: Advanced Find and Replace Questions

[Albus X]

I have officially submitted my proposal at
https://google-melange.appspot.com/gsoc/proposal/review/google/gsoc2013/albuseer/1

Please have a review.

On Mon, Apr 29, 2013 at 10:18 AM, Albus X  wrote:
>> I am not very familiar with git, what is the meaning of
>> [712e1871/lyxgit]? I just cloned the source now.
> I got it. It's the commit ID and I finally find out what your changed.
>
> On Mon, Apr 29, 2013 at 8:49 AM, Albus X  wrote:
>> On Mon, Apr 29, 2013 at 7:15 AM, Tommaso Cucinotta  wrote:
>>> Interestingly, those few scenarios which are documented in the manual, are
>>> not really included in the automated test cases as from 
>>> devel/autotests/findadv-*.in.
>>
>> In fact, it's not in the manual. The manual said I can search by
>> regex, and gave that example, but not said about how to use the regex
>> box. Maybe should add some explanation to the manual?
>>
>>> these are all nice observations, but they have to be raised on the list with
>>> some different subject line, e.g., "Usability issues in math-writing" or 
>>> similar,
>>> otherwise you won't get the necessary attention by the right people.
>>
>> Thanks for your pointing out.
>>
>>> At least, now there's a regression test for this [712e1871/lyxgit].
>>
>> I am not very familiar with git, what is the meaning of
>> [712e1871/lyxgit]? I just cloned the source now.
>>
>> Thanks.

Re: GSoC: Advanced Find and Replace Questions

[Cyrille Artho]

Hi Albus,
Thank you for your proposal.
I am not very familiar with the project, but I see that you mention Prolog 
in your proposal. Are you planning on using Prolog? Do you have any ideas 
in how a similar algorithm could be implemented in C or C++ instead? (It 
would be good if a LyX user did not have to install Prolog as LyX currently 
does not depend on it.)


Albus X wrote:

I have officially submitted my proposal at
https://google-melange.appspot.com/gsoc/proposal/review/google/gsoc2013/albuseer/1

Please have a review.

On Mon, Apr 29, 2013 at 10:18 AM, Albus X  wrote:

I am not very familiar with git, what is the meaning of
[712e1871/lyxgit]? I just cloned the source now.

I got it. It's the commit ID and I finally find out what your changed.

On Mon, Apr 29, 2013 at 8:49 AM, Albus X  wrote:

On Mon, Apr 29, 2013 at 7:15 AM, Tommaso Cucinotta  wrote:

Interestingly, those few scenarios which are documented in the manual, are
not really included in the automated test cases as from 
devel/autotests/findadv-*.in.


In fact, it's not in the manual. The manual said I can search by
regex, and gave that example, but not said about how to use the regex
box. Maybe should add some explanation to the manual?


these are all nice observations, but they have to be raised on the list with
some different subject line, e.g., "Usability issues in math-writing" or 
similar,
otherwise you won't get the necessary attention by the right people.


Thanks for your pointing out.


At least, now there's a regression test for this [712e1871/lyxgit].


I am not very familiar with git, what is the meaning of
[712e1871/lyxgit]? I just cloned the source now.

Thanks.


--
Regards,
Cyrille Artho - http://artho.com/
Until Eve arrived, this was a man's world.
-- Richard Armour

Re: GSoC: Advanced Find and Replace Questions

[Albus X]

No. I am indeed not familiar with programming in Prolog. I mentioned
it because the pattern matching feature of Prolog, (or maybe many
functional languages?) is implemented that way. I think find and
replace is also a kind of pattern matching, and for the replace
feature, the substitution table calculated in that algorithm does the
exact job for group referencing in regexp.

In the last one year and more, I focused mainly on the topics of
compilers and programming language design. Therefore this project
immediately remind me of that algorithm. I have indeed an description
of the algorithm in my proposal, therefore it can be implemented
easily in C++. Maybe my description is not clear enough, I will add
more detail to it now.

On Tue, Apr 30, 2013 at 8:44 AM, Cyrille Artho  wrote:
> Hi Albus,
> Thank you for your proposal.
> I am not very familiar with the project, but I see that you mention Prolog
> in your proposal. Are you planning on using Prolog? Do you have any ideas in
> how a similar algorithm could be implemented in C or C++ instead? (It would
> be good if a LyX user did not have to install Prolog as LyX currently does
> not depend on it.)
>
>
> Albus X wrote:
>>
>> I have officially submitted my proposal at
>>
>> https://google-melange.appspot.com/gsoc/proposal/review/google/gsoc2013/albuseer/1
>>
>> Please have a review.
>>
>> On Mon, Apr 29, 2013 at 10:18 AM, Albus X  wrote:

 I am not very familiar with git, what is the meaning of
 [712e1871/lyxgit]? I just cloned the source now.
>>>
>>> I got it. It's the commit ID and I finally find out what your changed.
>>>
>>> On Mon, Apr 29, 2013 at 8:49 AM, Albus X  wrote:

 On Mon, Apr 29, 2013 at 7:15 AM, Tommaso Cucinotta 
 wrote:
>
> Interestingly, those few scenarios which are documented in the manual,
> are
> not really included in the automated test cases as from
> devel/autotests/findadv-*.in.


 In fact, it's not in the manual. The manual said I can search by
 regex, and gave that example, but not said about how to use the regex
 box. Maybe should add some explanation to the manual?

> these are all nice observations, but they have to be raised on the list
> with
> some different subject line, e.g., "Usability issues in math-writing"
> or similar,
> otherwise you won't get the necessary attention by the right people.


 Thanks for your pointing out.

> At least, now there's a regression test for this [712e1871/lyxgit].


 I am not very familiar with git, what is the meaning of
 [712e1871/lyxgit]? I just cloned the source now.

 Thanks.
>
>
> --
> Regards,
> Cyrille Artho - http://artho.com/
> Until Eve arrived, this was a man's world.
> -- Richard Armour

Re: GSoC: Advanced Find and Replace Questions

[Albus X]

My proposal is updated. I removed the mentioning of Prolog, made the
algorithm description clearer, and added more explanation.

On Tue, Apr 30, 2013 at 8:58 AM, Albus X  wrote:
> No. I am indeed not familiar with programming in Prolog. I mentioned
> it because the pattern matching feature of Prolog, (or maybe many
> functional languages?) is implemented that way. I think find and
> replace is also a kind of pattern matching, and for the replace
> feature, the substitution table calculated in that algorithm does the
> exact job for group referencing in regexp.
>
> In the last one year and more, I focused mainly on the topics of
> compilers and programming language design. Therefore this project
> immediately remind me of that algorithm. I have indeed an description
> of the algorithm in my proposal, therefore it can be implemented
> easily in C++. Maybe my description is not clear enough, I will add
> more detail to it now.
>
> On Tue, Apr 30, 2013 at 8:44 AM, Cyrille Artho  wrote:
>> Hi Albus,
>> Thank you for your proposal.
>> I am not very familiar with the project, but I see that you mention Prolog
>> in your proposal. Are you planning on using Prolog? Do you have any ideas in
>> how a similar algorithm could be implemented in C or C++ instead? (It would
>> be good if a LyX user did not have to install Prolog as LyX currently does
>> not depend on it.)
>>
>>
>> Albus X wrote:
>>>
>>> I have officially submitted my proposal at
>>>
>>> https://google-melange.appspot.com/gsoc/proposal/review/google/gsoc2013/albuseer/1
>>>
>>> Please have a review.
>>>
>>> On Mon, Apr 29, 2013 at 10:18 AM, Albus X  wrote:
>
> I am not very familiar with git, what is the meaning of
> [712e1871/lyxgit]? I just cloned the source now.

 I got it. It's the commit ID and I finally find out what your changed.

 On Mon, Apr 29, 2013 at 8:49 AM, Albus X  wrote:
>
> On Mon, Apr 29, 2013 at 7:15 AM, Tommaso Cucinotta 
> wrote:
>>
>> Interestingly, those few scenarios which are documented in the manual,
>> are
>> not really included in the automated test cases as from
>> devel/autotests/findadv-*.in.
>
>
> In fact, it's not in the manual. The manual said I can search by
> regex, and gave that example, but not said about how to use the regex
> box. Maybe should add some explanation to the manual?
>
>> these are all nice observations, but they have to be raised on the list
>> with
>> some different subject line, e.g., "Usability issues in math-writing"
>> or similar,
>> otherwise you won't get the necessary attention by the right people.
>
>
> Thanks for your pointing out.
>
>> At least, now there's a regression test for this [712e1871/lyxgit].
>
>
> I am not very familiar with git, what is the meaning of
> [712e1871/lyxgit]? I just cloned the source now.
>
> Thanks.
>>
>>
>> --
>> Regards,
>> Cyrille Artho - http://artho.com/
>> Until Eve arrived, this was a man's world.
>> -- Richard Armour

GSoC: Advanced Find and Replace Questions

[Albus X]

Hello everyone, I am interested in GSoC 2013 and particularly the
Advanced find and replace project.

Before I start, I had a try with the current version, and want to ask
some questions.

I read the corresponding sections in the user guide, and tried to find
some math using regular expressions. I noticed that, I must insert a
regex box first. But since there can be regex box inside math box or
vice versa, which is the correct way?

Because there are multiple ways to enter the same expression, when I
failed to find the expression, I would staring at it wondering whether
I entered it in the correct form, which is confusing.

I tried to find \frac{x^2}{1+x^2} in that manual by typing
\frac{.*}{1+x^2}, \frac{[.*]}{1+x^2} and [\frac{.*}{1+x^2}], where [ ]
is the regex box, but no one works. And then I tried to find
\frac{[.*]}{[.*]} and it also doesn't work.

I understand there are many problems therefore it needs a redo. But
can I assume that it's not a bug but my mistake?

Following are other problems (or maybe suggestions):

1. As a programmer, I just can't get rid of my habit of typing a space
between a variable name and an equal sign, which causes lyx exiting
math mode immediately. As I can type as many spaces as possible in
normal mode without effect, this behavior involves inconsistency.
Maybe switch the commit key to enter will be better? This problem also
exists in matrix, cases environment etc. where typing space should
result in jumping to the neighbouring cells, and Enter jumping to next
row. And when I typing space by mistake, backspace should intuitively
acts like undo.

2. After I typed \frac{ in math mode, it show the frame of a fraction,
but with a pair of { } in the numerator part, which should be eaten I
think.

3. For latex users, it seems more natural to input \section rather
than click mouse twice.

Thanks.

Lyx version: 2.0.3 under ubuntu 12.10

Re: GSoC: Advanced Find and Replace Questions

[Tommaso Cucinotta]

On 28/04/13 13:55, Albus X wrote:
 Hello everyone, I am interested in GSoC 2013 and particularly the
 Advanced find and replace project.

Hi there,

welcome to LyX!

First of all, let me redirect you to the information available at

  http://www.google-melange.com/gsoc/org/google/gsoc2013/lyx

where you can find how to apply to GSoC 2013 for a project related to LyX
(the application procedure is outlined there).

Some issues related to FR have been discussed for quite some time, not
only in relation to GSoC'13, but also other times in the past. Just spend
some time searching throughout the mailing list archives, please.

 I read the corresponding sections in the user guide, and tried to find
 some math using regular expressions. I noticed that, I must insert a
 regex box first. But since there can be regex box inside math box or
 vice versa, which is the correct way?

Interestingly, those few scenarios which are documented in the manual, are
not really included in the automated test cases as from 
devel/autotests/findadv-*.in.

It may be possible that some of those don't work any more, but I need to check
all of them again, as the manual contribution was written quite some time ago.

 Following are other problems (or maybe suggestions):

these are all nice observations, but they have to be raised on the list with
some different subject line, e.g., Usability issues in math-writing or 
similar,
otherwise you won't get the necessary attention by the right people.

Thanks,

T.

Re: GSoC: Advanced Find and Replace Questions

[Tommaso Cucinotta]

On 29/04/13 00:15, Tommaso Cucinotta wrote:
 I read the corresponding sections in the user guide, and tried to find
 some math using regular expressions. I noticed that, I must insert a
 regex box first. But since there can be regex box inside math box or
 vice versa, which is the correct way?
 
 Interestingly, those few scenarios which are documented in the manual, are
 not really included in the automated test cases as from 
 devel/autotests/findadv-*.in.

At least, now there's a regression test for this [712e1871/lyxgit].

I can confirm it currently fails when ignoring format (the default, test 
findadv-re-05),
whilst it succeeds when searching with format-awareness (test findadv-re-06).

This needs a quick fix for branch, if possible.

T.

GSoC: Advanced Find and Replace Questions

[Albus X]

Hello everyone, I am interested in GSoC 2013 and particularly the
"Advanced find and replace" project.

Before I start, I had a try with the current version, and want to ask
some questions.

I read the corresponding sections in the user guide, and tried to find
some math using regular expressions. I noticed that, I must insert a
regex box first. But since there can be regex box inside math box or
vice versa, which is the correct way?

Because there are multiple ways to enter the same expression, when I
failed to find the expression, I would staring at it wondering whether
I entered it in the correct form, which is confusing.

I tried to find \frac{x^2}{1+x^2} in that manual by typing
\frac{.*}{1+x^2}, \frac{[.*]}{1+x^2} and [\frac{.*}{1+x^2}], where [ ]
is the regex box, but no one works. And then I tried to find
\frac{[.*]}{[.*]} and it also doesn't work.

I understand there are many problems therefore it needs a redo. But
can I assume that it's not a bug but my mistake?

Following are other problems (or maybe suggestions):

1. As a programmer, I just can't get rid of my habit of typing a space
between a variable name and an equal sign, which causes lyx exiting
math mode immediately. As I can type as many spaces as possible in
normal mode without effect, this behavior involves inconsistency.
Maybe switch the commit key to enter will be better? This problem also
exists in matrix, cases environment etc. where typing space should
result in jumping to the neighbouring cells, and Enter jumping to next
row. And when I typing space by mistake, backspace should intuitively
acts like undo.

2. After I typed \frac{ in math mode, it show the frame of a fraction,
but with a pair of { } in the numerator part, which should be eaten I
think.

3. For latex users, it seems more "natural" to input \section rather
than click mouse twice.

Thanks.

Lyx version: 2.0.3 under ubuntu 12.10

Re: GSoC: Advanced Find and Replace Questions

[Tommaso Cucinotta]

On 28/04/13 13:55, Albus X wrote:
> Hello everyone, I am interested in GSoC 2013 and particularly the
> "Advanced find and replace" project.

Hi there,

welcome to LyX!

First of all, let me redirect you to the information available at

  http://www.google-melange.com/gsoc/org/google/gsoc2013/lyx

where you can find how to apply to GSoC 2013 for a project related to LyX
(the application procedure is outlined there).

Some issues related to F have been discussed for quite some time, not
only in relation to GSoC'13, but also other times in the past. Just spend
some time searching throughout the mailing list archives, please.

> I read the corresponding sections in the user guide, and tried to find
> some math using regular expressions. I noticed that, I must insert a
> regex box first. But since there can be regex box inside math box or
> vice versa, which is the correct way?

Interestingly, those few scenarios which are documented in the manual, are
not really included in the automated test cases as from 
devel/autotests/findadv-*.in.

It may be possible that some of those don't work any more, but I need to check
all of them again, as the manual contribution was written quite some time ago.

> Following are other problems (or maybe suggestions):

these are all nice observations, but they have to be raised on the list with
some different subject line, e.g., "Usability issues in math-writing" or 
similar,
otherwise you won't get the necessary attention by the right people.

Thanks,

T.

Re: GSoC: Advanced Find and Replace Questions

[Tommaso Cucinotta]

On 29/04/13 00:15, Tommaso Cucinotta wrote:
>> I read the corresponding sections in the user guide, and tried to find
>> some math using regular expressions. I noticed that, I must insert a
>> regex box first. But since there can be regex box inside math box or
>> vice versa, which is the correct way?
> 
> Interestingly, those few scenarios which are documented in the manual, are
> not really included in the automated test cases as from 
> devel/autotests/findadv-*.in.

At least, now there's a regression test for this [712e1871/lyxgit].

I can confirm it currently fails when ignoring format (the default, test 
findadv-re-05),
whilst it succeeds when searching with format-awareness (test findadv-re-06).

This needs a quick fix for branch, if possible.

T.